Good morning, everybody.
Good morning.
It's my pleasure to welcome you to Corning's investor event right here at the Stock Exchange. I'd like to also welcome those investors joining by webcast. Before we begin our formal presentations, I'd like to remind you that today's remarks contain forward-looking statements that fall within the meaning of the Private Securities Litigation Reform Act of 1995. These statements involve risks, uncertainties, and other factors that could cause actual results to differ materially. These factors are detailed in the company's financial reports. You should also note that we'll be discussing our results using core performance measures, unless we specifically indicate our comments relate to GAAP data. Our core performance measures are non-GAAP measures used by management to analyze the business. You can find a reconciliation of core results to comparable GAAP values on the Investor Relations section of our website at corning.com. We have a great agenda today.
Our CEO will kick off with details of, and an upgrade to, the Springboard Plan, as well as an overview of how we drive profitable, durable growth over the long term and create value for our shareholders. Next, you'll hear from our business leaders, who will provide you with information and updates on how their businesses help drive Springboard, as well as from our head of research about additional innovation and growth opportunities. Finally, our CFO will wrap up the formal remarks, putting a financial perspective on the day, and from there, we'll move to Q&A. For those of you that are here with us today, we will head out to the demo exhibits. I hope you enjoy the day. We look forward to engaging with you. Now I'll turn the podium over to our chairman and CEO, Wendell Weeks.
Thank you, Ann. Welcome, everyone. It is really great to have you here with us. We have got a lot of exciting news to share with you, even better. You are going to get the chance to see some of our key innovations driving our success. More importantly, you are going to get a chance to meet some of the people who are bringing it all to life. Everything we are talking about this morning is tied to our Springboard Plan. Today, we are announcing a significant upgrade to that plan. Before we upgrade it, let us quickly review the original Springboard Plan for context. This is our Springboard Incremental Run Rate chart. The starting point is our fourth quarter 2023 sales of $3.27 billion. This provides a base of a $13 billion annualized sales run rate from which to spring. The Y-axis represents incremental sales above our quarter four 2023 run rate.
The X-axis represents time, marking each year's fourth quarter run rate through 2028. Now let's fill in some numbers, starting with our original internal Springboard Plan. Our internal plan is the output of the strategic planning process that we run with each of our market access platforms. These are our actual business plans. We set our objectives and our compensation based upon those plans. When our businesses submit plans to corporate, they factor in a variety of probabilistic outcomes. They try to account for the known unknowns. The plans aim for a 70% confidence interval, which means that based on their analysis, there is a 70% chance that they will deliver sales greater than or equal to that number. That said, there is a lot to take away from this slide.
First, we have a significant sales opportunity: potential growth of $8 billion in annualized sales run rate by the end of 2028, with $5 billion by the end of 2026. Second, we expect growth across all of our market access platforms, driven by a combination of upward cyclical and secular trends. We translate our internal Springboard Plan into a higher confidence plan for our investors. To do that, first, we focused on the next three years: a $5 billion opportunity by the end of 2026. Second, we applied a corporate-level risk adjustment to arrive at our high confidence plan to help you make your investment decisions. At the corporate level, we seek to probabilistically adjust for factors like macroeconomic cycles, changes in government policy, and timing of multiple secular trends and our related innovations.
Now, this is how we formulated the high confidence Springboard Plan we shared with you last year to add more than $3 billion in annualized sales by the end of 2026. Now, it's important to note that we purposely drew this as a wedge. We weren't trying to guide every quarter for the next 12 quarters. It obviously won't be a straight line. We are also not dealing with a hockey stick. When we built the plan, we expected to see strong growth in the first year. We told you last year that we already had the required production capacity and technical capabilities in place to deliver the sales growth, and that the cost and capital were already reflected in our financials. We expect to deliver powerful incrementals. As a result, our profits should grow faster than sales. This translates into an improving return profile.
Our Springboard target is to improve our operating margin to 20% by the end of 2026. This leads to improving ROIC, higher EPS, and strengthening cash flow. As you can see, we are off to a great start. Our quarter four 2024 sales were $3.9 billion. In the first year of the plan, we added $2.4 billion to our annualized sales run rate from our Springboard base of quarter four 2023. On that strong sales growth, we demonstrated powerful incremental profit and cash flow. In quarter four 2024, versus our Springboard starting point of quarter four 2023, we grew sales 18%. We grew EPS at more than twice that rate or 46%. We expanded operating margin by 220 basis points to 18.5%.
We expanded ROIC 390 basis points, and we closed out a strong year of free cash flow generation, delivering $1.25 billion for full year 2024, up 42% versus the prior year. Now, let's turn to the highlights of what we're actually announcing today. Okay, we are raising our first quarter guidance. Given the performance in January and February and what we see in our order books for March, we now expect sales to exceed $3.6 billion, and we expect EPS to come in at the high end of the range. We are also upgrading our overall Springboard Plan for you today. First, we are upgrading our internal Springboard Plan by $1 billion to now add $6 billion in annualized sales run rate by the end of 2026.
We hit or exceeded critical milestones in display, optical communications, and solar, and you will hear much more detail from Mike and Hal on these milestones, and Ed will provide a more detailed review of the overall plan. Quite simply, our strategies are working, and our customers are loving our innovations. We are increasing our internal plan from $5 billion to $6 billion by the end of 2026. Now, we once again applied a corporate risk level adjustment to translate our internal plan into a high confidence plan to help you make your investment decisions. Based on that, we are also increasing our high confidence plan by $1 billion. Our high confidence plan is now to add more than $4 billion in annualized sales run rate by the end of 2026.
We have raised both our internal plan and our high confidence plan to $6 billion and $4 billion, respectively, of incremental run rate sales. We continue to expect to deliver powerful incremental profit and cash flow on that sales growth because we already have the required production capacity and technical capabilities in place, and the cost and capital are already reflected in our P&L. This translates into an improving return profile. We continue to expect operating margin to improve to 20% by the end of 2026. Therefore, our upgraded sales will deliver higher EPS, stronger cash flow, and improved ROIC profiles than we originally anticipated at the start of Springboard. In total, we're giving you a very significant increase to our plan. We're essentially upgrading all key financial metrics today.
At the end of our session this morning, Ed's going to share more detail on everything I've just walked you through with some broader financial insight and perspective. We're also announcing some great progress against critical milestones that underpin our Springboard Plan. In display, we told you we were increasing price in the back half of 2024 to maintain stable $US dollar net income in a weaker yen environment. We said we expect to deliver net income of $900 million-$950 million in 2025, and net income margin of 25% consistent with the last five years. Today, we have the benefit of two and a half months of the first quarter under our belt, and we are seeing this play out exactly as we expected in our financials. We are able to confirm that we are on track to deliver these targets.
In Optical Communications, we have a lot going on, and it's all good. We are succeeding in capturing the Gen AI opportunity. Last year, we introduced new products for inside GenAI data centers, and we expected rapid adoption of these products to drive sales growth in our $1.3 billion enterprise business at a 25% compound annual growth rate from 2023 to 2027. We've received a tremendous customer response to our innovations, significantly growing our enterprise sales to a record $2 billion in 2024. Customer adoption continues to track ahead of our expectations, and as a result, we are raising our 2023 to 2027 CAGR from 25% to 30%. We also announced an additional revenue stream last year when we introduced a set of innovations to interconnect AI data centers.
We shared that we had reached an agreement with Lumen Technologies to provide our new GenAI fiber and cable system that enables Lumen to fit anywhere from two to four times the amount of fiber into their existing conduit. The agreement reserved 10% of our global fiber capacity for 2025 and 2026. Today, we are announcing that we have fully commercialized this product set, and we now have three industry-leading customers adopting the technology. Our production is tripling every month this quarter, so this innovation is now turning into a revenue stream to make a positive difference in our financials this year. Turning to our fiber to the home business, we told you we expected our customers to complete drawing down inventory they built during the pandemic and to resume buying at their deployment rates during the Springboard planning cycle. We now believe that process is largely complete.
Now, this is an important mitone for us as the conditions are now in place for our fiber to the home business to spring back to growth later this year. Mike is going to share more on exactly what we are seeing on all of these topics, and then we'll show you what our optical innovation sets look like in real deployments. When we introduced Springboard, we told you we would launch a new market access platform. Later in the year, we identified that it would be solar. Today, we're officially launching our new solar market access platform that we expect to grow to a $2.5 billion revenue stream by 2028. We expect a positive incremental impact on Corning's sales, profits, and cash flow this year. We are commercializing our new Made in America ingot and wafer products this year, and we have committed customers.
Our production will come online in the back half of this year. You will hear much more from Hal. We obviously have a lot of exciting news to share today. We are making great progress across the company, and this is our third Springboard event so far, with many more to come. We try to focus each of these touchpoints on one or two businesses where we have significant news on milestones and their financial ramifications. Because Springboard is a mitone-based plan, our goal at each session is to highlight the milestones that are actually moving our overall pace. As we reach certain inflection points, it changes our overall likelihood of success, which means we usually provide you with an update to the overall plan at these events as well.
Now, we host these touchpoints in person because we want to get you in the same space with the innovations that are driving our progress. At Corning, we invent innovative products manufactured using our proprietary processes and equipment in factories that we entirely own and operate in close proximity to our customers all over the world. This is how we build true decades-long business franchises. When you see our competitive advantage up close, it's worth a thousand words. We want you to see where these product sets work within our customer systems, how they make our customers' products better. We want to give you the chance to interact with the terrific people who bring their passion and expertise to bear to bring it all to life. Because the best way to really understand us here at Corning is to see us living out our mission.
That mission is the foundation of our Springboard opportunity, and it is at the heart of our approach to driving profitable, durable growth. We're celebrating our 175th anniversary next year. Our mission is to drive another 175 years of life-changing innovation. It all comes down to how we invest, make decisions, and evolve around important secular trends. We begin with our focused and cohesive portfolio. We invest in three core technologies, four proprietary manufacturing and engineering platforms, serving today five long-standing market access platforms. We are the lowest-cost technology leader in each. Through our deep commitment to R&D, we're able to anticipate and evolve around important secular trends and apply our earned insights to start early on the relevant technology vectors. We focus our strategy on capturing synergies among our core capabilities and then applying them to create disruptive innovations.
Overall, our success translates into vibrant businesses that grow profitably for decades, and our approach has placed Corning at the center of secular trends touching many facets of daily life. We do it all profitably because we're able to leverage our investments by reapplying and repurposing our assets and insights across multiple opportunities and markets. The more value we add in each market, the greater we multiply the relevance of our capabilities. Today, you're going to see right up close what this strategy looks like in action. First up are Mike and Claudio, and I'm really, really excited for you to hear what they have to share.
The three of us have been working closely together for years on the more Corning innovations driving our outperformance today in Optical Communications and on the next generation of innovations that we expect to help maintain our strong momentum in the business for years to come. Now, it's my pleasure to introduce you to Mike O’Day.
Thank you, Wendell. Wow. It is an exciting time for optical communications. Our Springboard Plan doubles the sales run rate and more than doubles our profit by Q4 of 2026 as cyclical and secular trends converge to drive our growth profile. We are one year into the plan, and I am happy to share that we are seeing tremendous results. Even more exciting is that there is much more to come. The opportunities we see today are unlike anything I have experienced in my 26 years with optical communications. Relationships with our long-standing carrier and enterprise customers have never been stronger, and we are innovating at a rapid pace. Now, I am confident we will capture these opportunities because our unique capabilities are more relevant today than ever. We have drawn on our insights to drive innovations that create new-to-the-world solutions that our customers love. That is why we are winning.
That is why I could not be more excited about the future of optical communications. Today, we are going to dive into three growth areas where our innovation is accelerating our success. Let us get started. First, we are raising our 2023 to 2027 enterprise sales CAGR from 25% to 30% as we capture the GenAI secular trend inside hyperscale data centers. We are seeing remarkable response to our new innovations from our customers, leading to a record $2 billion in revenue last year for our enterprise business. Second, we are creating an attractive new revenue stream outside the data center. In quarter one, we began shipping to Lumen Technologies our new fiber and cable system for data center interconnects. Two more industry-leading customers are now adopting it as well.
Lastly, carriers have worked through the inventory that they built up during the pandemic, setting the stage for fiber to the home to spring back to growth in 2025. Now, let's cover each of these one at a time, starting with inside the data center. Here, our opportunity is immense, and it continues to be a perfect fit for Corning's innovation model. We think a picture is worth a thousand words, so let me show you how things are changing. Not that long ago, the cloud represented our optical data center business that sits within our enterprise segment. And fiber, depicted here in yellow, was critical to connect this CPU-based front-end network. pre-GenAI, our enterprise segment grew at single digits. Then along came GenAI, which requires a back-end network that runs graphical processing units, or GPUs. Now, what do you notice?
A lot more yellow, which means more fiber, and that's because the GPUs all must connect to one another. Here we see a switch rack supporting 32 GPU nodes like Hopper, which require four times more fiber than a cloud network. That sparked a dramatic increase in the opportunity for Corning, which led to our enterprise growth in 2024 to $2 billion. Now, 2025 becomes the year of 72 GPU AI nodes like Blackwell, which is a much more compute-intensive GPU. That requires even more fiber. These nodes increase the fiber in the switch racks by another factor of four, highlighting a massive opportunity for Corning. You'll get to see an example of a 72 GPU node switch rack like the one in this picture later today in our demo lab.
You will also get a glimpse into one of our future growth opportunities that I'll quickly preview next. What you see here is a typical 72 GPU AI node with a switch rack and six server racks. As you can see, fiber connects from the switch rack to each of these server racks, and that's where the optical signal ends today. Now, you're looking at the back of those same six server racks. What don't you see? Yellow. What you do see, however, are gray bundles of copper cable, mini gray bundles. In fact, there are 2 mi of copper cable in each one of these racks, totaling over 12 mi of copper in this example. Guess what? That's where yellow goes next. Later this morning, Claudio will unpack why this is happening. More yellow creates opportunities to solve scale and density challenges with our insights and innovations.
We started over five years ago to solve these two challenges that all major hyperscalers face as part of their GenAI requirements. Let's start with scale, as that can mean many things. Let me show you what it means to us inside the data center. Here's an example of the enormous scale of GenAI. Mark Zuckerberg recently posted this announcement of a 1 million GPU data center campus that Meta is building in Louisiana. In purple, the size of Meta's data center campus is overlaid right here on Manhattan. It's so big that it would stretch from 2nd to 85th Street and from 2nd to 9th Ave. He used purple. When we look at a data center like this, we see yellow because we know it will be full of fiber. How much fiber? 8 million mi.
Now, that's a lot of glass. For context, enough glass to connect all 5 million single-family homes lit up on this map of New York State. Let that sink in for a moment. When you put the same amount of fiber into that Meta data center, it all has to fit in that tiny circle with the black dot covering the part of Manhattan that you see today. I told you this example was to share what scale means to us, but it also highlights how important density is to fit all this fiber, 8 million mi in this case, into an AI data center. Here we see the evolution of fiber density requirements and switching racks amidst the rising number of GPUs for typical GenAI nodes, as the switch racks are where density matters most.
32 GPU nodes have been the primary technology, with Hopper-like architectures increasing fiber by four times versus front-end cloud architectures. Next, we know that 72 GPU nodes, like Blackwell, are becoming more common. That increases the fiber 16 times versus traditional cloud switch racks. With larger nodes to come, we're already working on the requirements for the next GPU set, which we believe will require another major increase in fiber. Yet the size of that rack is not changing. It means a lot more fiber in the same space, screaming for even more Corning innovation. To solve these GenAI density challenges for our customers, we reinvented everything over the past five years: smaller fibers, cables, connectors. We put it all together to create customized AI solutions for our customers.
Not only did we make the fiber smaller, but we also made it better, providing a distinct advantage over competitive products. It bends tighter, allowing it to fit into spaces 10 times smaller than standard fiber without compromising optical link performance. This matters as more fibers in a rack means tighter bends and lots of pathway congestion. What you see on the screen is a great example of us delivering the right innovations at the right time. Our new components integrated into a customized high fiber count assembly, all done in our factories, so the customer installation is faster and error-free, all while capturing more value for Corning. Now, when we spoke to you in September, we described how we had created these innovations to solve the industry's density challenges.
What we couldn't tell you at the time is that these innovations were aimed at a 72 GPU AI node, like Blackwell, and these are the solutions that we have scaled and are selling right now to the biggest deployers of GenAI around the globe. As I close this section on GenAI inside the data center, let me tell you why our growth is sustainable. Corning has over 1,000 pending and granted patents worldwide relating to data center applications, with over 100 filed just in the last 12 months, highlighting just how fast we are innovating. We are not just inventors. We are makers. All these innovations we just talked about are made in the USA, and we are the largest, lowest-cost, vertically integrated manufacturer in the world. We leverage that to meet growing customer demand at scale.
Now, let's look at another way GenAI is fueling our growth, but this time with our carrier business. As we've worked on GenAI solutions for inside the data center, we saw that it would require a lot more data centers with much larger footprints. We also saw a need for significant growth in data transmission over long-haul networks to support GenAI training and inference. That prompted us to investigate how well these existing networks could support escalating AI traffic. By the start of 2023, we had our answer. Most routes were approaching their maximum data rate capacity, creating a need for many new high-bandwidth, low-latency links between cities and data center campuses. This essentially required a rebuild of long-haul networks. Think of this segment as data center interconnect, or DCI. Here is what makes this a major opportunity for us.
First, existing hubs require more fiber connections given the AI traffic growth. This is why Lumen contracted this for 10% of our fiber capacity last year. Next, GenAI requires the construction of up to 70 new hyperscale data centers each year in the U.S. alone, creating the need for more hubs. Meanwhile, power and land constraints are driving these new hubs to new regions. The U.S. currently has six major data center hubs today, and we anticipate that number could grow by more than double by 2030. Finally, all these existing and new data center regions need to be interconnected with fiber. Now, we were the first to tell you about this emerging need for high-density connections between data centers when we announced our Lumen deal last summer.
Since then, these network buildouts are moving faster, faster than expected, and the rest of the industry is realizing it as well. Let's look at what some of our customers and industry leaders are saying. Dycom shares how data centers are creating the highest level of interest in high-capacity, low-latency intercity networks in 25 years. Sienna projects substantial growth. It is clear there is strong industry belief of a growing opportunity for DCI. Density, as it turns out, is every bit as critical outside the data center as it is inside. Whether installing fiber in existing ducts or new ducts, our customers want to maximize the fiber capacity in the duct space since adding more fiber comes at a modest incremental first install cost.
For new DCI routes, micro ducts, like you see in this picture, are becoming our customers' preferred solution due to faster deployments, lower costs, and upgrade flexibility. With legacy cables, you can fit around 3,000 fibers in a seven-way microduct system, but our customers wanted us to do more to handle the escalating data demands of new AI applications. We willingly accepted the challenge to create a denser solution, took our core innovations from the inside data center, and applied them to this outside plant density challenge. This resulted in DCI solutions that leverage our new high-density Contour Flow cable and accommodate twice the fiber as legacy cables in the same diameter and duct size. Our customers love it.
Next, the fiber's patented optical features and proprietary manufacturing processes enable superior bend performance and protection against microbending that could occur as you double the density of fibers in the same space. In short, not only did we make it denser, we made it better. As a result of this innovation, we were able to increase the capacity of the trending seven-way microduct system from 3,000 to over 6,000 fibers while maintaining the same cable diameter. Now, that's a more Corning story. Twice the fiber in the same footprint made possible by our innovations. We're already winning in this space. Our DCI products are now fully commercialized, tripling output every month this quarter, and we've gained adoption of these products with three industry-leading customers, including Lumen as part of the deal that we announced again last summer. We're not done.
Looking ahead, we're working with these same data center operators to evaluate the benefits of new fiber technologies to address even greater density requirements and lower latency DCI links, gaining insights that will feed our innovation machine for future generations of products. Watch this space as we'll be back more with news soon. Now, let's look at our fiber to the home business, where we told you we expected our customers to finish drawing down inventory they built during the pandemic and to resume buying at their deployment rates during the Springboard timeframe. We think this graph really shows the point well. Here you see Corning's fiber shipments in relation to the market's long-term trend line. We first showed you this in the third quarter of 2023. We expected our customers to complete drawing down inventory they built during the pandemic, and now that process has largely concluded.
The conditions are in place for our fiber to the home business to spring back to growth. This is an important milestone indicating the stage is set for a carrier recovery later this year. We're also seeing confirmatory announcements not only from carriers like AT&T, but also healthy guides from major network installers such as Dycom. Now, once again, density matters here, and we've innovated the fiber cable connectors and terminals, delivering denser solutions that allow our customers to reuse existing infrastructure like poles and pedestals that accelerate their ability to pass more homes faster and at a lower cost. For rural deployments, we've cracked the code on how to help our customers get to these locations with more fiber, more affordably. Through trials with a major U.S. carrier, our new rural solutions have demonstrated cost savings of more than 30% for home passed.
Because all these fiber and cables are made in the USA, we expect continued strong adoption as government-funded deployments get underway in 2026. I hope you can see why I'm so excited about the opportunities before us as we apply our insight and innovation and our manufacturing capabilities right here in the U.S. to the secular trend of GenAI and the critical work of connecting the unconnected. When you think about GenAI, I hope you come away with a sense of how relevant Corning's capabilities are to the most important technology trend of our time. We're proud to be doing our small part to make that a reality. Why does Corning win? As we innovate, we draw on our three core technologies, four manufacturing and engineering platforms, and five market access platforms. Here's what that looks like for Optical Communications.
Fifty-five years ago, we ushered in a telecom revolution with our invention of the first low-loss optical fiber. From that moment, we've continued to extend our leadership, constantly winning new customers that grow into trust-based relationships that last for decades. Next, those customers bring us their toughest technology challenges, and we go to work, drawing on our world-leading expertise in glass and ceramic science and optical physics to develop new innovations that solve our customers' problems. Lastly, we bring these innovations to life through our low-cost manufacturing leadership, drawing on our capabilities in vapor deposition for fiber, extrusion for cable, and precision forming for connectors. No one else can do what we can do. That's why, after more than 50 years of industry leadership, the industry's biggest players continue to look to Corning to help build their next-generation optical networks. Next, you'll hear from Dr.
Claudio Mazzali, Corning's Vice President of Global Research, as he shares more about our work converting copper to fiber. Later in the demo lab, my team and I will help you experience these innovations for yourself. Thank you again for the opportunity to speak with you today. Claudio.
Thanks, Mike, and good morning, everybody. I'm Claudio Mazzali, and today I'll go a bit deeper on technology because it's really exciting to see how the three core technologies and four manufacturing platforms that Wendell described apply even more strongly to the next frontier in optical. Our full capabilities, including the fusion platform now, are getting pulled into the next innovation cycle. That is more Corning. I may be biased, but as a technologist, I believe the opportunities in front of us are even more exciting than what we have accomplished so far. Let me tell you why.
You just heard from Mike about the tremendous growth that our optical business is seeing today. To be ready to intercept the GenAI opportunity, we started working on these innovations several years ago. The results of that work are showcased here in the image on the left, an ocean of yellow cables connecting switches and server racks inside Hypers and GenAI data centers. That is the front-end or scale-out network. What is not visible in this picture is a back-end or a scale-up network connecting all GPUs within a GenAI node. There are a lot of connections on the back of the server racks connecting each GPU to one another. Converting that back-end to optical is a large innovation opportunity for us, potentially even bigger than the front-end network. To demonstrate my point further, let's look at the logical architecture.
On the top of this diagram, you can see all of the yellow links that Mike talked about. The front-end network interconnects all servers and switch racks to each other, and this is already all done optically. Now, let's turn one of the server racks around and see what's happening inside. You see this huge bundle of copper cables on the back of the rack? Those are the connections on the back-end network, which logically looks similar to the front-end, but with even more connection points. Today, there are more than 1,200 links connecting more than 70 GPUs using more than two mi of copper cables in a single server rack. Let me repeat, single server rack. This space is currently dominated by very short copper interconnects and can be another multi-billion-dollar market transition to optical. Now, I believe this transition is inevitable. Let me show you why.
For those that have followed us for a long time, this is how we think about the optical sector. We have been using the E-to-O framework to predict when different applications move from electrical to optical due to technical economic reasons. This framework uses the product of distance and data rate to define when it's more economical to be in the optical domain. That transition usually happens when the product reaches roughly 100 Gb meters per second, noted by the diagonal line here in the chart. Applications such as submarine, long-haul, metro, and even access have already transitioned to optical. The driver for this change was mostly cost to compensate signal loss and copper links. This all looks obvious now, but in my own career, I've seen this crossover happening in both metro and access.
Now, I'm not as old as Wendell, who probably saw that even in long-haul decades ago. The timing was always difficult to predict, but the transitions to optical were inevitable. Next were connections between data center buildings, Campus DCI that Mike described, and Intra DC. Data rates have been increasing from megabit per second to gigabit per second, and both spaces transition from copper links to optical as well. Once again, these are the links that Mike was talking about earlier. Now, we see both Hyperscale and GenAI data centers with architectures that are creating new requirements where electrical transmission over copper will be challenged, and we believe the transition to optical will be inevitable again. For instance, that server rack I showed a minute ago can contain a full GenAI node like Blackwell, and the entire node is contained in one server rack.
The back-end network is within the rack and can use more than 2 mi of copper per rack, as we said. You'll be able to see for yourself in the demo area outside. With data rates, data rates may still grow, the number of GPUs per node will also grow. There are architectures considered a single node with hundreds of GPUs distributed over more than 10 racks. Once that happens, the interconnection between GPUs on the same node will need to be extended to more than 30 meters. At that point, it becomes more cost-effective and energy-efficient to transition to optical. That is an inevitable transition. We are working with customers and partners specifically on this future evolution of AI nodes, but due to confidentiality agreements, I cannot share more specific details about what we're doing with them.
What we can do is share a little bit of what's on the public domain. Here's an example. This image is from a Meta public document, which shows that the compute fabric today is interconnected with direct attached copper or DAC, represented here by black lines on the left rack diagram. It's important to note that the signal turns to optical only at the network switch in the purple box in the middle. On the right rack, you see the cluster fabric of tomorrow, where all the copper black lines are now replaced by blue arrows representing optical links. In addition, the copper interconnects inside their servers and switches, represented by red lines on the left rack, are also replaced by optical links in their future architecture. An architect with much more optical links demonstrated the inevitable growth in optical that I described earlier.
Let's shift back to our capabilities. We are already bringing the optical signal all the way to the face plate, as you saw. We are working on helping the industry to bring that signal directly to the actual chip in the most effective way. It's time to go inside the box with Co-packaged Optics. Just look at this picture. Optical fibers are coming directly to a substrate holding an optical circuit board. From a technology perspective, fundamentals of this challenge combine unique requirements on thermal stability, optics, mechanical reliability. That was a hint that our materials could be useful here. Once again, we leverage our 345 way of reusing our technology, repurposing our assets, and reapplying our talents in a new area. Our Optical Communications sector, as Mike described, was already leveraging glass science, optical physics, and many of our manufacturing platforms for a long time.
Now we'll be also activating another capability not yet leveraged by the optical team, the fusion platform, as flat specialty glass could create highly stable substrates, which are more robust to high power densities than organic materials and offer the right composition and surface attributes for optical circuits. In addition to that, technologies to embed wave guides, both laser-written or actually using our proprietary ion exchange process, which is the same ion exchange technology we have been perfecting as a strength mechanism for Gorilla Glass. Finally, through glass vias, TGVs, alignment fiduciaries, advanced laser processes for simulation are all critical elements for us as we enter the CPO space. It is definitely in our space. Let's go quickly over why CPO has been getting so much attention lately.
As advanced chips get to 100 Tb per second of aggregated I/O, we started to have hundreds of lanes at more than 100 Gb per second each. Carrying those signals over a distance of just a few centimeters of copper requires advanced DSP, digital signal processing, or a stronger FEC, forward error correction. These advanced DSP and FEC not only add cost, latency, and consume power, they also waste valuable real estate of silicon that now has to be dedicated to the I/O section of the chip. Now, let's take a look at how the signal gets there and how CPO can help. As you can see here, the optical signal is converted to an electrical signal at a pluggable transceiver that's sitting at the face plate. After that, the signal must propagate over lossy copper lines until it gets into the I/O modules of the chips.
The ultra-high data rate propagation through lossy copper is what makes that advanced DSP and FEC a requirement. Now, with Co-packaged Optics, that large pluggable transceiver is eliminated and replaced by an integrated transceiver that is co-packaged with the chip. The optical signal now goes through a passive connector and directly to the integrated transceiver. Because the transceiver is so close to the chip and the lossy copper link is eliminated, most of that internal I/O can be dramatically simplified, recovering significant silicon real estate for data processing. In summary, CPO has great potential to save silicon real estate, power, latency, and cost, which are all very important elements for the scalability of Hypers and GenAI systems.
If we take this to the full vision, we can have a piece of flat glass becoming the entire substrate, which would reduce thermal issues with organic materials that we see today. Additionally, it would improve multi-chip packaging and enable an integrated optical circuit board with embedded wave guides and through glass vias. While there is lots of work to make it happen, none of those technologies are totally new to us. They are connected to the previous utilizations of our 345 cycle. This is a great example of how technologies developed for different Corning maps can complement each other in this new application. Optical connectivity and precision glass coming together and enabling us to penetrate a $40 billion semiconductor packaging market. Now, let's take a look on a short animation that will help make this connection clearer.
We start with a pristine sheet of glass from our fusion platform. Then we add optical wave guides using lasers or our ion exchange technology. Next come through glass vias and optical cavities, metalization and redistribution layers, and alignment features and fiducials for all the components that we have there. We use our proprietary laser technology for precise singulation, leaving edges with optical quality. After that comes the connectors and then the photonic integrated circuits, which are then ready for a broadband connection to switch ASICs or GPUs. Finally, at the end, you have an integrated optical circuit built into a glass substrate with integrated connectivity for ultra-high capacity chips. You're going to be able to see this demo exactly as you see here outside. We have been working with customers, and you may see demos from others as well.
Now, as Wendell keeps reminding us, material science is slow. We must start well before a secular trend becomes obvious if we want to have a chance to intersect it. In the same way, early insights and customer connections got us prepared a few years ago to intersect the density challenge that we are seeing now. The understanding of off-chip connectivity gave us the hunch to start looking into that problem. Believe it or not, what you see on the left of this slide is from a video shared at our 2019 Investors' Day. While it was not a hot topic then, we knew that the distance between the optical signal and critical chips, such as switch ASICs and GPUs, had to be reduced. We also knew that organic substrate may reach a wow as data rate goes up and power dissipation increases.
We knew that density of optical interconnects would need a new framework, dramatically different than traditional pluggables and connectors. Maybe it was not inevitable, but we started looking to it. Today, the same overall CPO concept of our Copackage Optics is a critical element on the roadmaps of many key players. Some industry estimations indicate that CPO can deliver more than 50% in power savings, more than 40% in cost reduction in optics, and roughly four times increase in density versus incumbent technology. That early work got us prepared to be in a leadership position, working directly with partners that are planning to implement these concepts on their next generations of high-capacity chip interconnects. We have joint papers, demos, even product announcements with key players in this space. We are working under NDA with many of them.
While we are still in the beginning phases, we are excited with how well our core capabilities align with this particular area. I recommend that you check out the CPO demonstration outside. Finally, Ed asked me to explain a little bit of what we have been doing on CPO and beyond, but I also want to take this opportunity to share why it's a very exciting time to be working at the Corning Labs. We are paranoid about understanding our materials to the deepest levels, and we are humbled to accept that customers, at least initially, know more than us about applications and problems to be solved. We are always in listening mode. When that combined understanding has an obvious connection with our 345, things get really interesting. What's ahead for us? A lot.
Here's a peek on some of the areas that I am personally really excited about. To start, the architecture space is one of the largest glass markets in the world. Given new regulations, we have an opportunity to enable advanced windows that are lighter and with improved energy efficiency versus today's designs. Augmented reality is also taking us to a different space in glass formulation, where we're achieving optical attributes well beyond what we thought was possible. Another major trend we're looking at is energy storage. Our work on a new-to-the-world process to make continuous thin ceramic separators has got some of our partners and potential future customers really excited. Bendable glass is a natural in our portfolio, thickness between 10-20% of a business card, and it's already in use in some current applications.
Once we get to that level of control in our precision forming, it will open other very interesting applications beyond consumer electronics. Carbon capture, a nascent technology, is still being proven out. From our perspective, it is a great example of leveraging our 345 approach. We do not need to invest much to become a significant player if this approach to help our planet proves to be essential. Finally, solar is a fantastic space to leverage both optical science and optical physics. You will hear more about solar from Wendell and how in a minute. In addition to what they will cover, we are also working on some interesting ways where our materials can help with both efficiency and reliability of solar panels. Yes, I am definitely biased, and I am sorry if I may sound overexcited, but it is just a fun time to be working for Corning Labs now.
With that, I would like to turn the podium back to Wendell. Thank you.
That's part of my job, is I get to work with people like that every day. Claudio just shared how excited he is about solar in our innovation portfolio. Now, we have been deeply interested in that space for well over a decade. Let me tell you why. Here is what a typical solar panel actually looks like. Now, it's basically a glass sandwich that has an anti-reflective coating and a polysilicon semiconductor material in the middle. Solar is fundamentally about converting the energy of photons into the energy of electrons within that material. Now, when we look at that system, what we see is that 80% of those poor photons are being wasted and that the overall conversion efficiency of that system has been stuck at around 20% for decades.
We see an area that is ripe for innovation. We also see that 90% of the mass in that solar panel is materials in which we have adjacent world-class capabilities. We make the best technical glass in the world. We apply coatings to our strength and vapor deposition and have a long-standing breathing position in polysilicon for semiconductor materials. I mean, not only is this an opportunity ripe for innovation, but it is also right in our wheelhouse. We have been seeking a low-risk, high-return entry into the solar industry. In 2020, we found our entry point. We acquired full strategic control of Hemlock Semiconductor, which had manufacturing assets with a replacement value of over $4 billion that were largely idle for less than $0.10 on the dollar.
We believed we could fill those assets by bringing Corning's unique abilities in material science, manufacturing, and engineering to serve what we saw as the emergent need for American-made products in the critical semiconductor and solar industries. Since acquiring these assets, we made process advancements to improve the purity of our polysilicon material, allowing us to increase our market position and to serve the highest-end chips in the industry. As a result, we are on track to double our semiconductor business by the end of the decade. Starting in 2020, the first Trump administration's focus on U.S. content gave us an advantaged re-entry into the solar market and the ability to restart our idled capacity. More recently, we added the capability to turn our advantaged domestic polysilicon into higher-value wafers, all made on the same campus. We are fully commercializing our wafer products later this year.
Now, for the entire platform, we now have committed customers for 100% of our capacity available in 2025 and 80% of our planned capacity for the next five years. We expect the solar business to incrementally add to Corning's sales, profits, and cash flow by the end of this year. We have been able to do all of this with customer funding and government support while generating positive cash flow every year. I want to illustrate our low-risk entry approach to solar in numbers. We have taken an asset where we had a minority ownership in Hemlock Semiconductor that was generating about $50 million a year in cash flow for us in the form of distributions before 2020. We have now turned that into a platform with $1 billion in sales at profitability levels above the Corning average.
Now, this chart shows how we have generated over $1 billion in cumulative cash flow from 2020 to 2024 with positive cash flow in every year. We expect another year of positive cash flow in 2025, and our cash flow will continue to grow as we build our $2.5 billion solar map. We found our low-risk, high-return entry point for Corning and its shareholders. Now, I'll turn it over to Hal to share our plan.
Thanks, Wendell. I'm glad to be here today to tell you more about an exciting and significant opportunity that we have in solar. As you heard, we found our low-risk, high-return entry point to build a powerful solar market access platform. We have already turned those assets into a platform with $1 billion in sales.
Today, I'm here to tell you how we intend to accelerate our growth in this map, delivering $2.5 billion in sales with above Corning average profit margin that will generate billions of dollars in cash. We're already well on our way, expecting to add to Corning's sales, profits, and cash flows this year. We're doing this using the building blocks that Wendell laid out. We made process advancements to serve a higher-end chip segment in semiconductors. We activated idle assets to serve the need for domestic solar polysilicon and added capability to transform our polysilicon into higher-value domestically made solar wafers, all vertically integrated together on our campus in Michigan. With these actions and assets in place, we are well positioned to more than double our revenue by 2028, as we've seen strong customer pull for our American-made products in the critical semiconductor and solar industries.
In fact, we are sold out this year and already have customer commitments for more than 80% of our capacity for the next five years. Together, these opportunities that we are executing upon support the U.S. goals to ensure independence in the semiconductor and energy sectors, both of which are critical to the U.S. economy and national economic security. The opportunities in these industries are rooted in silicon, which we know extremely well from our history. Silicon is a part of a core set of elements that we have deep knowledge of. When we combine silicon with other elements in our core, we form the basis of our innovations and the major products that all of us know that you know us all for, from glass and ceramics to now semiconductor materials. We're using an element that we know.
It's right in our wheelhouse and leveraging our core capabilities and assets to meet a need in a growing market. As you know, silicon forms the foundation of modern microelectronics. Given its clear advantages relative to other materials, it now accounts for nearly all semiconductor as well as solar wafer production. What we're doing with silicon in these two industries is a classic more Corning opportunity. Let's start with semiconductors, where we were one of only two companies in the world to first produce polysilicon, launching this industry and the technology revolution that followed. For over 60 years, we've been manufacturing successive generations of hyper-pure polysilicon to serve the advancements in the semiconductor industry.
High-quality semiconductor-grade polysilicon is one of the purest man-made substances on Earth, typically produced at extreme purity levels, which in the industry is known as 12N or 12 nines, as you see here on the slide. To put that in perspective, this is equivalent to one part per trillion, or said another way, finding one marble on the entire island of Manhattan. This level of purity is very difficult to achieve, and therefore, only a few companies in the world have the capability to make and supply this critical material that is at the foundation of the integrated circuit. Given our leadership in this space, our hyper-pure polysilicon is found in nearly every electronic device in the world, from cars and phones to appliances and automated industrial applications. As the only U.S. company with this capability, we enable the U.S.
To achieve strategic control at the front of the semiconductor supply chain. With our insight in this industry, we saw there was an opportunity to grow our semiconductor business as chip manufacturers began advancing their designs. We made process advancements in our polysilicon production, driving improvements in core and surface purity for the next generation chips. These upgrades have allowed us to sell into a more premium segment, essentially doubling our revenue within the decade as the industry pursues more advanced chip designs. We now have long-term customer commitments in place through the end of the decade, utilizing our existing capacity to enable our customers' innovation and growth. There is additional more Corning opportunity as customers continue to adopt even more advanced chips and national security interest in domestic chip manufacturing continues.
We also re-entered the solar market, where polysilicon plays a vital role as the foundational material for producing solar modules. In fact, demand for solar polysilicon has outpaced demand for semiconductor polysilicon, both globally and in the U.S., but we were not participating in a significant way. Let us take a look at what is driving the opportunity in the U.S. market. The U.S. market is the world's second-largest solar market, growing rapidly at 30% CAGR. This growth has been market-driven, as solar is now one of the country's lowest levelized costs of electricity for utility-scale buildouts. Further growth is anticipated in solar, as energy demand is expected to increase to support data centers and GenAI infrastructure. Solar's favorable economics make it an important part of utility providers' portfolios to serve growing energy demand. Today, solar is a significant contributor to the nation's newly built capacity for utility-scale electricity generation.
This trend is expected to continue, given its favorable economics and significantly faster deployment times than traditional energy sources to meet the accelerating energy demand in the U.S. To serve this growing demand, the solar market is dominated by high-performing crystalline silicon technology, which is highly dependent on foreign imports. That supply chain is shown here. As a result, over the last decade, the U.S. government has taken a series of policy actions, including the implementation of polysilicon sourcing requirements to support the rapid deployment of solar and drive a domestic supply chain to ensure U.S. energy independence. With these policy actions, it's no surprise that we're experiencing strong demand for our solar products, as we are one of the few companies able to meet the U.S. requirements for polysilicon and domestic sourcing. In fact, you don't see any domestic supply at all for wafers.
To enable domestic sourcing and serve growing demand, we took actions to both activate idle assets and move downstream into wafers, all from our state-of-the-art campus in Michigan. Activating idle polysilicon assets allowed us to increase the supply of our advantaged domestic polysilicon to meet the growing need for American-made products. Given our strategic source of polysilicon, as well as strong customer relationships, we saw a more Corning opportunity by converting a portion of our polysilicon into wafers with high value, moving downstream to establish a domestic source of wafers that did not exist before. As a first mover in establishing a domestic supply chain, our high-quality silicon-based wafers will enable today's most advanced solar cell technology with high conversion efficiency. With wafers, we are creating an exciting opportunity where we will capture more than two times the value of our polysilicon.
To serve the wafer opportunity, we expanded our Michigan campus with a new facility. This is the largest and first of its kind in the country. We've been making great progress in standing up this plant. We have some recent footage to share with you and give you a better sense of where we are. As you can see, we are making great progress. To stand that plant up in less than a year, that is something that only Corning can do. I hope you'll stop by our demo area and learn a little bit more about what we will be doing on the inside of that plant and see our products. Startup of our operations is on track, and we are set to begin shipping wafers in the second half of this year.
Across the solar map, we now have committed customers for 100% of our capacity this year and more than 80% of our capacity for the next five years. Let me give you an example from one of our customers. You may have seen a recent announcement about a partnership with Heliene and Suniva that demonstrates what our wafer capability is enabling for a domestic supply chain. This partnership will bring the market the only solar modules made with polysilicon wafers and cells manufactured right here in the U.S., an exciting step for the domestic supply chain. By expanding our position as a domestic supplier, we are serving growing energy demand and building a domestic supply chain that is critical to U.S. energy independence. Corning is ideally suited for this.
We have deep expertise in materials conversion, the ability to scale advanced manufacturing processes, and a proven track record in low-cost manufacturing. Leveraging our technical expertise and manufacturing capabilities, we will also continue to explore more Corning opportunities moving forward. Across the solar supply chain, we see the potential to capture even more value and improve performance and energy conversion. We believe this is really just the beginning for innovation in this space. With our strategic polysilicon assets already in place, we are now vertically integrated through solar wafers and uniquely positioned to meet the need for domestic content. As I mentioned, we are sold out this year and already have long-term customer commitments in place. This is a spectacular opportunity for us, creating significant near-term value from our solar map, delivering a positive impact on Corning sales, profits, and cash flow this year.
By the end of 2028, we plan to grow revenue to $2.5 billion, more than doubling our $1 billion business. All in, the map will deliver above Corning's average profit margin and generate billions of dollars of cash. We are at the start of what we see as a decades-long profitable business for Corning. Needless to say, we'll have additional milestones to reach on our journey in building out this business, and we'll plan to update you on those in the coming months. We look forward to hosting you at our Michigan campus soon so you can see the progress firsthand. With that, I'll turn things over to Ed to talk more about Springboard. Thank you.
All right. Thank you, Hal. Good morning, everyone.
It's great to be here today and to see so many familiar faces and some new ones that are new to our story at a very exciting time for the company. As you heard throughout the morning, we're reaching several exciting milestones in our Springboard Plan, with many more yet to come. I'm going to sum up the day, and I'll tie everything together from a financial perspective. In particular, I want to address a few topics that I know are on your mind. I know that the external environment is a focal point right now, so I'll share how we're thinking about tariffs and the macroeconomy. We upgraded both our internal and our high-confidence plans, so I'll spend some time on our approach for our corporate risk adjustment. We expect to generate significant cash over the Springboard timeframe, so I'll share our capital allocation priorities.
I am looking forward to opening things up for Q&A so we can hear what else is on your mind. I will start with our first quarter outlook. The guidance we provided on our January earnings call was sales of $3.6 billion and EPS in the range of $0.48-$0.52. Given the performance we have seen through February and what we have in our order books for March, we now expect sales to exceed $3.6 billion, and we expect EPS to come in at the high end of the range. We are off to a great start. Our innovations are working, and we expect to grow from these levels as the year goes on. In general, we are thrilled about our Springboard opportunity, the results we have delivered, and the upgrade we have provided. We are all aware that this is happening at a very interesting time.
For example, tariffs are a hot topic in our investor discussions. Let me share a little bit about how we're thinking about them. First, let's start with the direct impact to Corning from tariffs. Our company has a long-standing philosophy to locate our manufacturing operations close to our customers. We find that geographic proximity leads to better innovation and more delighted customers. This also has the benefit of serving as a natural hedge against global trade tensions and tariff structures. We apply this philosophy globally. Take the United States, for example. We have a large advanced manufacturing footprint in the U.S. This includes our Optical Communications business, where we have the largest fiber factory in the world in North Carolina.
We also manufacture products for our Environmental Technologies business, our Life Sciences business, our Solar business, and we even make products for our Mobile Consumer Electronics business in the U.S. We have a very strong manufacturing program here. Almost all of the products we sell in the U.S. originate from our manufacturing facilities in the U.S. In fact, nearly 90% of our U.S. revenue comes from products of U.S. origin. The majority of the remaining 10% is generated from products that are fully compliant with U.S. MCA rules. Of course, it is harder to predict the potential indirect impact of tariffs, for example, exactly how our customers could be affected. What we have done with our Springboard Plan is to try and factor macroeconomic risks of this nature into our corporate-level risk adjustment. We will talk about our risk adjustment a little bit more in a moment.
That's an example of what takes us from the $6 billion internal plan to the $4 billion high-confidence plan. Our Q1 guidance rates also factor in our best thinking on tariffs. With that in mind, why don't we get more into detail on our Springboard plan? As you heard from Wendell today, we are providing you with a compelling upgrade to our Springboard plan. First, we're raising our internal Springboard plan by $1 billion from $5 billion to $6 billion of incremental annualized sales by the end of 2026. When we say internal plan, we mean the output of our strategic planning process we run with each of our businesses. We set objectives and compensation based on those plans. Each business aims for a 70% confidence level, which means there's a 70% chance their plan will be greater than or equal to that number.
We sum all of those plans up, and that's how we get to the $6 billion opportunity by 2026. We have several springs activating now in display, in optical, in solar, and others that we expect to activate as we go forward. For example, in Gorilla Glass, we continue to innovate to add value and gain more dollars per device. We expect our triple-digit auto glass business to triple sales from the beginning of 2024 through the end of 2026. We focused our event today on our opportunities through 2026 because we have such significant growth over the next two years. We wanted to provide our investors with a closer look at what is driving that growth in the near term because it's happening so fast. Of course, the same secular drivers will also drive growth in the longer term.
You heard some great examples today from Claudio, Mike, and Hal. In particular, Claudio and Mike shared the opportunity in front of us in co-packaged optics. Our approach has been to hold Springboard events when we have significant news to share on milestones and their financial ramifications, like the milestones we are sharing today. We are going to continue to do that as we go forward. Now, let's talk about how we translate our $6 billion internal plan into a $4 billion high-confidence plan. We follow the same process for our upgrade as we did when we launched Springboard. We applied a corporate-level risk adjustment. At the corporate level, we seek to probabilistically adjust for factors like macroeconomic cycles, changes in government policy, and timing of multiple secular trends and our related innovations, fundamentally those factors that are not necessarily included in our business's strategic plans.
For example, I shared with you how we factor in forces like the direct impact of tariffs on Corning. We also seek to consider the indirect impact on our customers as part of our corporate risk adjustment. Another great example would be in our new solar market access platform. Now, you just heard from Hal we plan to grow our solar business to $2.5 billion of sales with profitability above the Corning average by 2028. That is what Hal is on the hook to deliver. In addition to the risk that Hal builds into his plan, we apply a corporate risk adjustment to go from our $6 billion internal plan to our $4 billion high-confidence plan. Our risk adjustment includes our probabilistic assessment on the potential for any changes in government policy.
Now, those are just some examples of how we adjust our internal plan to our high-confidence plan for investors. More importantly, both our high-confidence plan and our internal plan assume a significant increase in sales from our Q1 2025 levels. Our plans imply between a 20% and a 30% increase in sales within only about a year and a half. We already have the required production capacity and technical capabilities in place to deliver the sales growth. The cost and capital are already reflected in our financials, so we expect our profit to grow faster than sales. Our upgrades clearly illustrate that we expect even more robust sales growth than we initially laid out last year. We set an operating margin target of 20% about six months ago, and we made nice progress.
In the fourth quarter of 2023, the starting point for Springboard, our operating margin was 16.3%. In the fourth quarter of 2024, just one year later, it was 18.5%, a 220 basis point improvement. We remain confident we will deliver the 20% target by the end of 2026, and there may be upside depending on how some of our milestones play out. Of course, the higher sales, profit, and cash flow we expect lead to an improving return profile. With that, let me take a moment to talk about some of our Springboard milestones we've achieved. In display, we are maintaining our market technology and cost leadership while benefiting from market growth and a glass supply demand environment that is increasingly balanced to tight. We successfully implemented double-digit price increases in the second half of 2024 to ensure we can maintain stable U.S.
Dollar net income in a weaker yen environment. We also hedged our yen exposure for 2025 and 2026 with hedges in place beyond 2026. In 2025, we reset our yen core rate to 120 yen to the dollar, consistent to our hedge rate. We are not recasting our 2024 financials because we expect to maintain the same profitability in Display at the new core rate. Now, in the first quarter, we're seeing this play out exactly as we expected in our financials, and we remain confident that we can deliver net income of $900 million-$950 million in 2025 and net income margin of 25%, consistent with the last five years. Our first quarter results will provide the first proof point that we are on track to deliver. This is reflected in the updated guidance I discussed a moment ago.
Now, you just heard from Mike a lot of exciting things in the optical space. I'm not going to repeat the same details because you're about to see the products and meet the people bringing it all to life. I do want to make a few points. We're raising our expected enterprise sales growth rate for the period 2023 through 2027 from the 25% CAGR we provided you last year. The rapid adoption of our new products for inside GenAI data centers is driving the increase. Continued strong performance in this business factors into our updated first quarter guidance. In our carrier business, our new GenAI innovations for data center interconnect have been fully commercialized, and we now have three industry-leading customers adopting the technology. Additionally, in carrier, we have a cyclical spring that is just beginning to activate.
We told you that we expected our customers to complete drawing down inventory they built up during the pandemic, primarily for fiber to the home, and we expected them to resume buying at their deployment rates. We now believe that process is largely complete. Conditions are in place for our fiber to the home business to spring back to growth later this year. Beyond that, we expect growth tied to government initiatives to connect the unconnected starting in 2026. That is another example of the external dynamics we account for in our corporate risk adjustment. Overall, we're very excited about the growth we're pursuing in optical. I like how Wendell summed it up earlier. We have a lot going on, and it's all good. Now I'll turn to solar.
Our solar business is currently part of our Hemlock and Emerging Innovations Group, which is designed to launch early-stage projects and growth opportunities such as our auto glass business. I want to build on what we've shared about growing the new solar map into a $2.5 billion business by 2028. We found a low-risk entry point into the solar market by gaining full strategic control of Hemlock in 2020. This came with solar polysilicon assets that were largely idled. Essentially, we acquired a set of assets for $0.10 on the dollar. The replacement value of those assets would be about $4 billion.
As you heard from Wendell, we took a business where we had a minority ownership and were receiving about $50 million a year in cash flow in the form of distributions and turned it into almost a $1 billion revenue business at profitability levels above the Corning average. Additionally, we generated over $1 billion in cash from 2020 through 2024, and we expect 2025 to be another year of positive cash flow. We have been able to do all of this with customer funding and government support, all while generating positive cash flow every year. We made process advancements to serve a higher-end chip segment in semiconductors. We activated idled assets to serve the need for domestic polysilicon and added capability to transform our polysilicon into higher-value domestically made solar wafers, all integrated together on our campus in Michigan.
We have committed customers for 100% of our capacity available in 2025 and 80% of our capacity for the next five years. In addition to the risk Hal is building into his internal plan, we apply a corporate risk adjustment to go from our $6 billion internal plan to our $4 billion high-confidence plan. Our risk adjustment includes our probabilistic assessment on the potential for any changes in government policy. Even adjusting for our probabilistic outcomes in our corporate risk adjustment, our high-confidence plan assumes we add incremental sales, profit, and cash flow later this year. With that, let me shift to capital allocation. The upgrades to our internal and high-confidence plans include higher sales and higher profit. We expect to convert that higher profit into more cash flow.
As an early proof point, in year one of the plan, we grew free cash flow 42% for the full year versus the prior year. Always front and center to us is capital allocation. How do we choose to invest the expected higher cash flow? Companies do capital allocation in different ways. We prioritize investing for organic growth opportunities that drive significant returns. We grow primarily through innovation. We believe this creates the most value for our shareholders over the long term. Our investors have confirmed this. As we see high return opportunities in the future, we will invest in those opportunities. Wendell described what seemed to our investors like overnight successes, such as our new products for GenAI. The truth is we were investing five years earlier in the innovations that would enable the strong results that we are seeing right now in optical.
The reason we were so successful is that we were ready for the inflection. We will maintain that approach with our innovations for GenAI, for example. The co-packaged optics opportunity that Claudio shared is another great example of where we are investing right now. While that growth opportunity is likely beyond the 2026 timeframe, we invest in the near term to ensure robust growth for the long term. We have led in optical communications since we invented the first low-loss fiber 50 years ago. As each connection has moved from copper to optical, we have led the way with the best products for each application. Our leadership has resulted in decades-long franchises and financial annuities for our investors. We also seek to maintain a strong and efficient balance sheet. We are in great shape. We have one of the longest debt tenors in the S&P 500.
Our current average debt maturity is about 23 years, with only about $1.2 billion in debt coming due over the next five years. We have no significant debt coming due in any given year. Finally, we expect to continue our strong track record of returning excess cash to shareholders. We already have a strong dividend. Therefore, as we go forward, our primary vehicle for returning cash to shareholders will be share buybacks. We have an excellent track record. Over about the last decade, we've repurchased 800 million shares, or close to a 50% reduction in our outstanding shares, which at today's share price has created about $19 billion in value for our shareholders. Because of our growing confidence in Springboard, we started to buy back shares in the second quarter of 2024, and we have continued to do so since then.
In the first quarter of 2025, we will invest another $100 million in share repurchases as we continue to believe in investing in ourselves. Thank you for joining us today. When we originally launched Springboard in the fourth quarter of 2023, we provided you with a compelling financial plan. In the first year of the plan, we delivered compelling results. From our Springboard starting point, we grew sales 18%. We grew EPS more than twice that rate, or 46%. We expanded operating margin by 220 basis points to 18.5%. We expanded ROIC by 390 basis points, and we generated $1.25 billion of free cash flow for the full year, up 42% versus the prior year. As we begin year two, we just gave you an even more compelling plan with our upgrade.
We look forward to seeing you at future Springboard events where we expect to once again share even more compelling results. Thank you for joining us today. Now I am happy to open things up for Q&A, so I will turn things back to Ann.
You need water? Yes. Thank you. Okay, I have a couple of folks that have microphones so that everybody on the webcast can hear your question. Raise your hand and we will get going. I see Steve Fox first right here in the middle. Hang on a second. We will get that mic to you.
Thanks. Thanks for having me and thanks for all the great detail. I was wondering, given all the financial color you provided, if you could zoom in a little bit more on the optical operating margins. Last year, they went from 13.8% in Q1 to 18.3%.
It seems like you're introducing more innovation that should have more value. Can you talk about how much operating leverage is left to get out of optical and how much innovation can maybe drive a better mix for the business? One quick follow-up. Ed, I just want to make sure I understand the free cash flow comments going forward. How would free cash flow growth in the next year or two compare to operating profit growth, better or worse than the growth in profits? Thanks.
Thanks. You handle both, brother.
All right. I would say on the first one, we certainly can continue to expand our net income margin in optical as they continue to grow. I think as you grow in a business, sometimes you have some cost to ramp for new capacity or production.
We have experienced that a little bit as we grew in 2024, but I think there is more room to run on the margin on optical. Yes, I expect our cash flow to continue to grow from 2024 levels. We talked a little bit about CapEx in January, so I think our 2025 CapEx will be similar to our depreciation rate. I think that is a good metric to use. Because we are not spending a lot on capital, cash flow should expand around in line with operating margin. I do not want to provide too specific a guidance, but I think that is a decent rule of thumb. As our operating profit grows, our cash flow should grow.
Talk about the innovation part of that answer on margin. Are you getting better pricing or better returns on your innovation in optical than maybe three years ago? Thanks.
Yes.
It's an important part of the operating leverage that you're seeing. Those products that you saw from Mike and that you will see in the demo and the generation that comes after that, you're seeing a continued narrowing of the amount of people in the world that can do it. And what you saw from Mike is the enormous amount of intellectual property now getting filed at a faster and faster rate. Those things usually come together to provide the opportunity for premium pricing. Then it's a matter of getting all the cost structure right because some of these products, they're not easy to make, brother.
Okay. We'll come up here to Tim Long.
Hi. Thank you, Tim Long. Barclays here. Also, too, if I could, maybe first on the discussion about the scale-up architecture, copper to optical transition.
I think last year, OFC, one of the industry players said there's a 10 to 1 CapEx difference if you make that switch today. If you could just talk a little bit about that chart you have with the price gap there. Where is that current gap and how quickly do you think that can close? I'm sure there's other things other than just acquisition costs. Second, maybe Claudio, for you, I don't know if you're going to take the first, but I think Ed said no major CPO opportunity until after 2026. Curious your view of timing and what you think the intermediary steps might be, LPO and opportunities along the way until we get to more of a full CPO architecture. Thank you.
I'll take the first.
You want to take the first one?
Sure.
On the first one with regard to the cost of CPO, maybe I'll answer the question this way. We are excited about the opportunity with CPO, and what's happening in the space right now is a lot of copper that we'll show you in our demo in just a moment, connecting the GPUs in the back of the server rack. We do that at a certain cost per gigabit, call it $0.05 per gigabit. What's happening as more GPUs go into that node, into that server rack node, and as the applications for AI increase, more data has to be used to compute these large training models. As a result, you need more GPUs. More GPUs creates more power and more bandwidth consumed in that rack.
As a result, that's where the opportunity for optical comes into play because more data at higher data rates requires an optical link versus a copper link, and it reduces the power by a factor of half, plus or minus, inside that node. The challenge for us is to continue to come down the cost curve on the optical piece to get in line as much as we can with the price per gigabit or the cost per gigabit of a copper link. That's what our innovations are aimed at doing right now.
Yeah. On your question on timing for CPO, as you saw here, right, we try to work, we try to answer several questions, right? We try to first decide, does the fundamentals make sense? Right? Is this going the direction to solve a real problem?
We look into can we apply our capabilities to actually help on that problem? When those things get together, we say, okay, now let's talk to customers. Let's talk to partners and make sure that they are in the same align with us. Those three things are happening with CPO right now, right? The fundamentals, yes, solve a problem. Our capabilities apply well for that. Customers are asking us for help. That's why we're working on this. You mentioned LPO, which is linear pluggable optics. It's basically when I mentioned here removing the digital signal processing, the DSP or FEC from the transceiver, there are ways to actually try to remove that and make a pluggable transceiver that's simpler. That's the LPO that was mentioned on the question. To me, that just validates that it's a problem.
The cost of DSP, the energy that is creating is a problem. Now, we have been replacing copper for decades, and it is always like that, right? There is a problem, and you work on an ultimate solution for that. Until you get to that point, there will be intermediate approaches. To me, that is just a validation that is a real problem to work on. I think DSP will be used. I am sorry, LPO, as you mentioned, linear pluggable optics, will be used short term. I do not think that changed the big challenge that we are going. CPO will come.
I believe that we're going to see, depending on, as Mike said, depending on the data rates, depending on the distance, depending on the size of the GenAI nodes that different players will play with, some of them will start with LPO pluggables, some of them may be moving to CPO sooner.
I want to make one more comment, if I could add one thing. For me, I think my job is to make sure we have a good high confidence plan. That's why I made that comment that we think CPO is largely outside the 2026 window. I think the most important thing is getting the innovation right. The timing will be the timing, right, as the industry evolves. For us, we have to get the innovation right.
I think that's what we're doing, and that's what Claudio and his team are doing, Mike and his team are doing. That's how we think about it. That's why I wanted to make that particular point because we spend so much time on it.
We're going to be wrong on time, but being early is better than being late, so.
Exactly.
That's all, Samik.
Hi, Samik from JPMorgan. You talked about Wendell in your prepared remarks, the two new data center interconnect opportunities outside of Lumen that you've sort of now started supplying to. Can you talk about the relative size or magnitude of the opportunities there relative to you've already been quite detailed about the Lumen opportunity? Maybe if you can share a bit more details of how you're thinking about that. Staying within optical, how are you treating BEAD within the Springboard plan?
Are you changing sort of your timelines or your magnitude of sort of contribution from BEAD? And a second one for Ed, you did mention you're trying to bake in some of the demand impact from the tariff potentially in some of the outlook that you're providing or the Springboard Plan. Maybe if you can flesh that out a bit in terms of which segments are you really thinking about the demand impact a bit more? Thank you.
Let's divide it this way. I'll let Mike address from your perspective in our internal plan, things like BEAD, etc. I'll just make one quick comment on DCI. So the size and scale of this opportunity is quite large. The question is one of timing.
What we're fundamentally looking at is a rebuild of the entire long-haul architecture, which Claudio in his own humorous way pointed out that perhaps I was old enough to have seen that cross over the line. It is a pretty significant move. What you see us doing is approaching that pretty conservatively. We are getting a large number of inquiries to be able to do basically that build-out and not only do that build-out, but fill more conduits than we ever filled when we did long-haul. We're not comfortable enough yet to place those in our financial planning exercises. I think that could change if we roll ahead about six months. To save your question a little bit, I think we'll have some more news and then we'll feel more confident. Perhaps you'd address BEAD and the like.
Yeah.
Just to add on DCI and BEAD both, with DCI, the one interesting thing that's happening as we engage with customers, we talked about Lumen, but two others are adopting our technology, really driven by GenAI and the use in these data centers and the builds really from a different set of customers from the first long-haul build-out. These are data center operators wanting bespoke pathways, if you will, to connect a data center training center in one city with another city. That is what really is contributing to this encouragement that we have in DCI. With regard to BEAD, I would say from our perspective, there's nothing in 2025 in our financials that we expect from BEAD. It's 2026.
There are 35 or 36 states now that are in the final phase of the approval process with the government and NTIA that have gone through the entire process successfully. No money has flowed yet. I think with the administration's discussion about BEAD, how do we accelerate and get BEAD moving to enable and get more homes connected? We stand ready. I think that's a next-year thing. When money starts to flow, our capacity is in place in the U.S. as required to satisfy the demand for BEAD. We are watching closely and carefully. Starlink satellite remains a viable technology, especially in areas that are maybe harder to reach or more expensive to reach with fiber, building a wire line network. I fully expect that to have a spot as it has through the timeframe of BEAD as well.
We stand ready if and when BEAD money starts to flow, engage with all of the customers in the ecosystem as well.
Ed, do you want to pop in on tariffs? Second part of Samik's question.
On tariffs, yeah. I think maybe just stepping back on our risk adjustment again, I do not think you should think about it as we risk adjust very specifically for very specific things. We take our entire internal plan and we run a number of scenarios. We do a lot of modeling, and our goal is to, in the aggregate, get to a high confidence plan. That is the way we think about it. Certainly the macro environment factors into that quite a bit.
Tariffs for sure factors in, but I think I shared how tariffs impacts us directly, which would be the impact I think we would see in the very short term.
Great. Up here to Asiya.
Hey, thank you, Asiya from Citi. Ed, if I may, just on the capital allocation, like you listed out $100 million for the 1Q , why not lay out a longer-term plan given high free cash flow generation, high confidence plan? Seems like the risk adjustment factor from your original Springboard Plan is lower now, so you have more confidence in your non-risk adjusted plan than you did before. Why not lay out a little bit longer-term capital allocation framework? Should we expect more investments coming? That's always the top of mind for investors. Thank you.
Yeah, I think for me, the most important thing is we've achieved a number of milestones to date, but we still have many more milestones that we want to have in front of us to really size both the high confidence, the internal plan, not just for this time window, but beyond 2026 as well. We may lay out in the future a sort of more concrete capital allocation plan, but for today, we really wanted to focus on that short-term window and the growth opportunities we have in front of us. As I mentioned, our primary vehicle for our excess free cash flow is going to be buybacks. That's the way we want you to think about it leaving here today.
We hear your question as a suggestion, and we will take it appropriately. We'll be back to you.
Josh up here up front.
And then Matt. Josh, can you just—
Thanks. Josh Spector with UBS. I wanted to follow up on the solar opportunity that you guys highlighted. A couple of things. First, is any of this opportunity reliant on U.S. IRA staying intact as is? You talked about efficiency and the improvements that Corning can bring to that. What type of efficiencies are you looking at in terms of power conversion? Is that something that's being priced into some of the offtakes you have with customers?
Thanks. I'll take the first one. You take the second one.
Sure.
In regard to your first question about government incentives for the 45X, I think the way that I think about it is that, as you know, government plays a role in influencing energy markets all over the world, which is really why, as Wendell alluded to, we wanted to find that low-risk entry point that we did. That really started back during the Trump 1.0 administration, where we found our way in the low-risk way. We have built that billion-dollar platform to now. We have generated billions in cash, cumulative cash. That has made us be really well positioned to think about going forward. Now, going forward, I mean, humbly, I do not know what governments will do. I do not know what form they will take, whether they will be incentives, whether they will be tariffs, whether they will be subsidies.
What I do have a high confidence in is that this administration or any other will continue to put a priority on U.S. energy independence. With a priority on U.S. energy independence, one will see the desire for domestic sourcing and a domestic supply chain. That is where we come into play and can put ourselves to work. Of course, I feel like our customers are telling us the same thing. The reason they are telling us the same thing is because we are sold out this year for our domestic product. In fact, we have more than 80% of our capacity, as we have said a few times, spoken for for the next five years. We are in pretty deep discussions in regards to the remainder of that capacity. That is the plan that you heard me talk about that I intend to deliver on.
What you heard Ed say is that he's factoring in any sort of variability that we may see along the way. I think we feel really good about our position.
Efficiencies?
Yep. On the efficiencies, perhaps the easiest—I'm not going to even talk about the silicon itself, but let's think about the panel, right? I think the easiest way to understand why we are excited about that we may have some good ideas for that is that just look at the panel. Think Wendell showed the sandwich, right? You have a large piece of flat glass there. That sounds like us. That glass has to be super strong. Think about Gorilla Glass, what we have done there to make glass stronger. I do not want the photons to actually be reflected and going back. I want to use anti-reflective coatings, for instance.
We just launched anti-reflective coatings with consumer electronics that are unmatched by anybody else. You need those photons to go through that glass. You need a glass that can let the photons go through. Think about what we have done in optical fiber. That is all about the glass design to get photons going through. When you combine all those things that you need in a panel to the business that actually we have been working on for decades, we are confident that we have some ways to improve that, putting that all together into a panel in a way that perhaps people may have not thought about before.
Matt.
Hey, thanks. It is Matt McDonald from Deutsche Bank. Two follow-up questions. First on CapEx, maybe for Ed. You laid out the plan for $6 billion non-risk adjusted, $4 billion risk adjusted.
I just want to make sure, does that entail CapEx staying at current levels, low $1 billions? The second question, maybe for Wendell, you've got a pretty broad business. You talked to a lot of customers. What are you seeing and hearing from them in light of tariffs and maybe more of a ramp-up in macro volatility in the last couple of weeks?
Yeah. On the first question, I want to reiterate, we are investing. We've been investing. We've invested through this cycle. We haven't added significant capital or capacity over the last couple of years, and that's why our CapEx has come down. I think more or less we have the capacity in place to deliver the $6 billion, the upside range in the next two years. A lot depends on exactly where it comes from and how it comes.
Even beyond that, I think we can support continued growth with our current footprint. The good news, I think, would be if we're going to invest in capacity, that means we have a significant growth opportunity with a high return. We will do the best we can to de-risk that opportunity with customer funding and all the other tools we use when we add capacity. That's the way I think you should think about it. It doesn't mean we won't have capacity in the future. In fact, I think you want us to be adding capacity in the future. I also want everyone to take away that we are continuing to invest in opportunities through this cycle.
Great. Wendell?
What a fascinating question.
On tariff structures, I think the main source of strategic anxiety is the lack of clarity so far in what form it will take. Historically, tariff structures have not much impacted flows. It causes, let's say, for something like consumer electronics, last time we faced this, there was a substantial tariff structure put on televisions, for instance, from China. What that did was it repositioned relatively rapidly the supply chain so that those now flow through Mexico and through other countries. Therefore, because capitalism is fast and governments are slow, right, you tend to flow around whatever those barriers are. This time around, I think what is interesting is what form will it take and will those barriers be more universal or not. To the degree which there is arbitrage available, capitalism will get all over that at speed.
To the extent there is not, right, then you're just looking at elasticity of demand or how it moves the other countries' currencies given their relative elasticity. If you're some of these countries that have 2, 3x worldwide capacity of what demand is, pricing, what price you'll strike at becomes a relatively arcane exercise. I think in almost any space, business will find a way. The problem right now, we're just in one of these time periods where we don't know what it is yet. I must say that for us, there are times when it is really nice to be a boring values-based company who just sort of says, "Well, wherever our customers are, I think that's where we should make our product." We're enjoying that natural hedge at the moment. We'll still feel what happens with our customers. Does that make sense?
Great.
Up here in the second row.
Hi. Peter O'Keefe, O'Keefe Stevens Advisory. One of the areas we did not discuss much today was automotive and with the gas particulate filters, the Gorilla Glass in windshields, and bringing down the weight in automobiles and the automobile turning into an entertainment area as well. Can maybe talk about some of the exciting stuff that is going on in automotive because we only briefly scratched the surface on that today.
We will be back with on automotive, we will be back with another Springboard event. What we try to do with these events is touch on those areas where we are seeing really significant milestones. We were starting to have a lot of big news that would impact our financials in the near-term buildup. You are quite accurate.
Basically, what we're seeing in worldwide trends is actually a longer life too and more being asked of internal combustion engines, which bodes very well for our various environmental products. As well, the move towards more and more demand for entertainment, lightweighting those ICE vehicles, etc., as well as electric has sort of put us in the spot where we're feeling really confident about our automotive opportunities. We'll be back and share all that with you.
Thanks.
This is [Saru] right here. Yep.
Hi. Hi. Sorry. Yep. Hi. This is [Saru] from Mizuho. For the new Springboard Plan, can you provide us some color on where the $1 billion incremental revenue is coming from across segments? The new Springboard Plan delivers EPS, free cash flow, and/or ROIC, but why hasn't the operating margin unchanged at 20%?
Yeah. I'll start.
I think the biggest areas of change maybe here in the short term have been in optical and in solar, as you heard today. That is why we wanted to share our thoughts. Our sort of confidence in the near-term has increased, and our opportunity in terms of size has increased as well. Certainly, there are other things that are positive. I think on the margin target, I made the comment in my remarks, and hopefully, it was clear that I do think there is opportunity for us to be above the 20%. I think we want to actually get closer to that 20% before we raise that target. We made a significant leap forward in 2024, over 200 basis points of improvement. I think we will continue to accrete up towards that 20%.
At some point, if we have confidence or new milestones that we want to share, we'll update you on that target.
Anybody wants to add anything else?
Woo Jin.
Thanks. Woo Jin Ho from Bloomberg Intelligence. Just a question in terms of the incremental growth on the enterprise optical. How much of that is from greater rack density versus your line of sight to the AI data center buildout? I have a follow-up to that.
Yeah. I would say our 25%-30% raise in the enterprise CAGR is largely because of the density increasing in data centers.
As I mentioned, GPU nodes moving from existing and what drove our growth in 2024, particularly moving to the 72 GPU nodes or nodes like that where GPUs are increasing, drives the need for optical by another factor of four, four times more fiber over what we've seen in really the last year or so. That, as more bespoke GenAI data centers come online, higher GPUs in those data centers requires a lot more yellow, essentially. That is the basis for our enterprise raise.
Got it. In terms of the confidence of going to 30%, are there any capacity commitment metrics that you're able to share with the audience?
Sorry. Any capacity?
Yeah. Capacity commitments from the enterprise sector.
Yeah. We don't have necessarily any capacity commitments from our customers.
What we continue to do, I'll start with first, we have largely the fiber cable capacity in place to deliver the growth that's associated with the 30% CAGR. We are continuing to add in some small areas our capacity, particularly in the connectivity area, which is typically lower capital cost to get more connectivity products developed in our solutions and putting connectors on. You'll see that in our financials. It's all in our financial plan as part of our raise to the 30%. We are adding capacity quickly as we're getting more demand from our customers and more signals of increased demand from our customers.
I think if I could put you into a hypothetical conversation with one of our key customers, that would be typical. It would be, A, I want more, right?
B, tell me what I have to do to make sure that I can get more flexibly delivered where I want. I also want the ability to change my product design depending on what sort of racks am I plugging into, how am I building my clusters. It gets very difficult for them to actually do like we do on DCI, where we know the lengths and we know what we're trying to do. Here, you see a very aggressive set of ideas from our customers of what they want to do. I get power here. I get land grants here. Everything is constantly moving and shifting. Therefore, even sometimes what you'll hear from some suppliers that then causes a lot more concern is, "Wait a minute, this company is drawing back from certain of its commitments in certain countries," which impact that.
All that's really doing is shifting around where the clusters are getting built, what scale, and how. So far, almost every call Mike gets is a request for more. That does not mean that lasts forever. It is just our life right now. Is that a fair characterization?
Yes. That is right. Lots of uncertainty. We field lots of uncertainty and engagement with our customers. In general, most of the conversations, if not all of the conversations, are, "How can you get me more? And can you get me more with your capacity?" That is what we are trying to do. Great.
We have been sitting here for a long time. We have been talking about demos. I think it is time that we get up, stretch our legs, and go see the demos. Thank you all for joining us.
Thank you.