Hello, ladies and gentlemen. Thank you for joining us virtually today for our latest progress update, with this session focused on our network infrastructure business. I'm David Mulholland, Head of Investor Relations, and I'm delighted to have many of the members of our network infrastructure leadership team on the line with me. As a reminder, we host these sessions to provide more detail and color on the journey that we're on in each of our businesses. The presentation used in the event will be available on our IR events page on our IR website after the event today. Before we get started, a quick disclaimer. During this event, we will be making forward-looking statements regarding our future business and financial performance, and these statements are predictions that involve risks and uncertainties. Actual results may therefore differ materially from the results we currently expect.
Factors that could cause such differences can be both external as well as internal operating factors. We have identified such risks in the risk factors section of our annual report on Form 20-F, which is available on our investor relations website. In terms of the structure for today's event, I'll shortly hand over to Federico and then the unit leaders for their presentations. Then we'll have a short Q&A session. With that, let me hand over to Federico to start the presentation session.
The fifteen months since we last met have been tough for our industry, and in many ways, this is surprising. The component shortages and the resulting build-up of the customer inventory we've seen, as post-pandemic phenomena, geopolitics continue to make life challenging, but that's nothing new. Higher interest rates and inflation have, in many places, eased, and in others, have become part of the new normal. Nevertheless, these trends persist and have led to what Dell'Oro calls inventory correction. We call it drag on our business, and Nokia is far from immune to these trends. Looking ahead, we remain remarkably confident in our industry and our company. As customers spend down their inventory, they naturally look to the next building cycle. As a result, as Pekka and Marco have pointed out, in recent quarters, order trends have been improving in network infrastructure.
Today, I will outline some important trends that we believe are driving the next wave of investment and explain why network infrastructure is perfectly placed to catch that wave. When the network infrastructure was formed at the beginning of 2021, we were still in the grip of the pandemic, with people from office workers to TV presenters moving to virtual work. Demand for excellent home connectivity grew fast, but this did not create the wave of investment. It just accelerated a trend that was already there. Investment in networks was already being driven by the replacement of copper cabling with fiber, worldwide 5G rollouts, and the growth of the cloud, and of course, it was not a matter of chance that we could take advantage of that demand. Smart investments had ensured we were in the right place at the right time.
Our Quillion chipset meant we had a 25 Gb PON product very early. Of course, these days, we have demonstrated 50 and 100 Gb. Our FP4 chipset has seen more than 300 deployments by the end of 2020 and was joined right on time by the next generation, FP5, in 2021. We caught up with in our optical portfolio with PSE-5, going to market in 2021, and then the launch of PSE-6s last spring gave us a technology advantage over our rivals. All of this contributed to a string of great quarters for network infrastructure. But as we have fought with these, the industry-wide challenges I have identified, we've seen for the first time year-on-year revenue declines. Nevertheless, the fundamentals of our business remain strong.
All our competitors have struggled to some extent over the last year, and we can see that Nokia has maintained, and in some cases, even increased its position. Analysts place our share at number one in edge routing and number two in total routing globally, excluding China. In XGS-PON, ONT as well as G.fast, we are number one globally. In Optical Networks, we are number three globally, and we are number two in APAC, excluding China, and number two in Europe. In short, we have gone through the recent challenging quarters in good shape, and we are ready to take advantage of the opportunities the future holds. Now, looking to the second half of the decade, several trends are shaping our markets. I would like to outline them for you briefly here before the division heads go into more detail in their upcoming presentations.
The first trend is the drive for connectivity. As Sandy will explain, there is still headroom in the fiber market. Fiber now passes more than half of households worldwide, but the number connected lags behind, and even in richer nations, there are pockets of unserved and underserved communities. Meanwhile, pressure on networks is leading to upgrades for higher speeds. The second trend is increasing interest in security. When almost every device in your house can have an IP address, the number of devices that can participate in a distributed denial-of-service attack expands exponentially, but it's not just DDoS that is a threat. We're also looking ahead to topics such as quantum computing... The third trend is geopolitics, which touches on a range of areas, including supply chain. The fourth trend is scale, but with efficiency. Particularly at times of economic uncertainty and challenge, this topic is essential for our customers.
And finally, the fifth trend is the widespread adoption of artificial intelligence, which has implications for our industry, of course. In responding to these trends, we ensure that everything we do, everything, is aimed at helping customers either derive value from their network investments, helping them achieve lower total cost of ownership, or doing both. In short, we are in business to help our customers make money and save money, and we do that while aiming to be the most trusted partner in our market. Our first priority is to grow our customers' network and their business. This is an important distinction. Growing networks is inescapable. Even companies that think they cannot invest have to invest. That's driving continued FTTH business, both to premises not yet passed by fiber and to those passed but not yet connected.
But we also must help our customers grow their business, whichever business they are in. Growth in broadband continues, as we can see from the extremely exciting announcement AT&T made just today, which Sandy is going to talk more about it in a moment. Broadband, based on our technology leadership and our strong customer focus. But it's not just fiber rollouts that are continuing at a pace. In fixed wireless access, for example, we are making an entirely new market. CSPs have invested enormous amounts in rolling out 5G. Now, we're helping them get value from those investment by using fixed wireless access to connect underserved communities. Incidentally, the ambition to connect the world is one we're proud to have, and it is becoming more realistic with government stimulus programs.
We are excited that Nokia was able to take the lead in readiness for the bid program in the U.S. Sandy will talk more about all these topics in a moment. Meanwhile, our Multi-Access Gateway or MAG, delivers more capacity with a lower total cost of ownership. That's why it's been chosen by T-Mobile in the U.S. and Du in the Middle East. But new customers include a leading Indian operator. That shows that our solutions are applicable in global market without looking at the geography, in very different economic geographies and organization patterns. The home is a major bottleneck in broadband services, so our customers are expanding their in-building business with our Wi-Fi solution and our Corteca applications. Finally, our customers are able to gain more value through slicing, which we offer in fixed, of course, IP and Optical Networks.
Enabling customers to divide up the network for enterprise, wholesalers, backhaul services, and so on, is a value-added services for their subscribers, making it another way for them to grow their business. With threats from distributed denial of service attacks, increasing networks need faster and more accurate DDoS mitigation, something that Deepfield Defender is fully equipped to do. In combination with the 7,750 routers, Deepfield helps customer protect against attacks from both outside and inside the network. While DDoS attacks are happening right now, we tend to think of the security threat from quantum computers as a problem for later. However, bad actors using the harvest now to use later techniques are occurring and a real threat today. Luckily, Nokia's technology can protect today against tomorrow's dangers.
This is built in, not bolt on, to both our IP networks and Optical Networks portfolio, making it cheaper as well as more effective, and it's unique to Nokia. We are already using this in Europe and elsewhere as a selling proposition, particularly for enterprise customers. Geopolitics is a major theme in our markets and will continue to be influential in the remainder of the decade. While government policy in some areas is certainly a factor, we also see customer demand for so-called clean networks in several geographies. It's a contributory factor in a number of our wins. For example, we've recently closed a big optical deal in Europe, partly based on Nokia's reputation for high ethical standards and behavior. At the same time, interest is growing in digital bridges. Subsea cables are generally considered very secure. However, security concerns have recently been triggered.
First, subsea bridges, most notably, ironically, the Nord Stream pipeline, which isn't a telecom cable, but it got high visibility. Second, the long-running regional conflict in the Middle East has rendered the Red Sea unsafe for shipping. As a result, we see some customers wanting to use a combination of subsea and long-haul terrestrial networks to form digital bridges. This gives us a wider range of routes and increases redundancy. Until recently, the efficiency battleground has been networking technology, and this very much continues. Successive silicon designs in NI have achieved greater scale and high efficiency, including in power consumption. Now, however, there is a new arena for efficiency battles, and that's automation. Our NSP WaveSuite and Multi Planner solutions provide automation to help customers reduce OpEx and shorten service delivery times. But there are other drivers for automation that are coming in surprising areas.
One of the biggest problems our customers face is the retirement wave hitting their staff. Newly qualified engineers are ready for tomorrow's technology, while today's networks use today's technology, so maintaining those networks become more and more difficult. It's a problem with a novel solution in the shape of AI. Our SR Linux platforms are already allowing people to communicate with network elements in natural language, and that is helping bridge the generation gap. Of course, Nokia is not the only company that is using AI. According to McKinsey, technology based on generative AI will add more than three percentage points annually to global productivity growth. Two-thirds of respondents to a survey by the consultancy said their company is regularly using the technology, AI. At the same time, as often happens with newly emerging technologies, there is a backlash underway.
The recent stock market correction in U.S. stocks is just one sign. My team is largely unfazed by all this. Just as we don't get caught up in the hype, so we don't despair when the promised land turns out to be a little farther away. Whether AI is a revolution or an evolution, it will surely be the defining technology trend of the decade. Indeed, a new study by our colleagues at Nokia Bell Labs has found that AI will contribute a quarter of all WAN traffic by 2030. This is growing to almost 600 EB a month. Increasing use of artificial intelligence tools drives an increase in the need for compute, in turn, creating pressure in data centers.
It is in the area of data center, that is connectivity to and from the data center, and connectivity inside the data center itself, that we see the biggest opportunity for growth. Success in this market, of course, means succeeding with the large web-scale companies who are becoming the most significant new network builders. But we also see considerable opportunity for our traditional CSP customers in selling to and partnering with web scalers, and we are actively participating with them on opportunities to do so. For example, e& in the Middle East chose to go with Nokia's interconnect solution to provide FP5-based connectivity services to hyperscalers. Meanwhile, Telekom Malaysia has chosen to use our Optical Network solutions to cope with hyperscalers' connectivity demands.
These deals are good examples of how telco operators will respond to the future and of the importance of innovation and performance to our entire ecosystem. The two major strategic moves we announced at the end of Q2 both contributed strongly to our future ambition. The planned divestment of ASN helps to create increased focus in our three core businesses, while giving our submarine networks business a secure future. The sale of the business will decrease our revenue temporarily, but will boost our margins, and that will give us more scope for investment. When it comes to our proposed acquisition of Infinera, the deal is highly complementary from a geographic viewpoint, and Infinera's customer base in the U.S. is balanced by our strength in APAC, Europe, Middle East and Africa, and Latin America.
The acquisition also strengthens our relations with web-scale, as well as boosting our focus on the optical data center interconnect market. It is too early to discuss the future shape of the combined portfolio, but conversations my team and I have had with customers, including major hyperscalers, have been really positive. They are excited about the scale and reach of the deal and what it would bring to the combined organization. James will say much more about this in his presentation. The proposed Infinera acquisition is focused on increasing the scale in our optical business, but we are leveraging the full network infrastructure portfolio to meet the needs of the new network builders. In our IP network business, our solution in IP data center networking and data center fabric combines advanced technology with customer-focused functionality.
Our data center networking solution includes the 7750 Service Router, which we launched almost a year ago, as well as the existing 7250 IXR series. The 7750 Service Router is based on FP5 chipsets to provide the industry's first 800GE network solution, supporting ever-increasing data center connectivity growth. Our intra data center switching fabric includes next-generation data center switches based on the SR Linux network operating system, providing an open, automated, and reliable system. But we'll talk much more later about the IP network business, including a trailer for some upcoming developments. We have also signed deals and agreements, some public, some confidential, with customers, including Tencent Cloud targeting enterprise customers in APAC, Equinix, which chose us for both inter-data center, intra-data center connectivity solution, and others in several geographies....
We have also won the technical phase in some important bids and landed frame contracts with customer groups, including hyperscale, web scale, and enterprise customers in the research and educational network and healthcare sectors. Indeed, data center networking is one of the most important growth engines for network infrastructure enterprise segment. In this, as in other areas, partners play a strong role in our plans. I am delighted that Kyndryl, one of the world's largest global system integrators, has adopted our data center networking portfolio to support its Fortune 1000 customer base. We are steering this partnership through our enterprise and partner success team, which was formed just at the beginning of the year. The team is making great progress, building on the double-digit CAGR delivered by network infrastructure and enterprise growth in recent years.
Our strategy is to focus on mission-critical networks for transportation, energy, and public sectors, among other market segments. We also see opportunities to partner with CSPs, tower companies, fiber companies, and others to expand into the digital infrastructure and network-neutral host areas. Recent analyst assessment of 2024, in some respect, makes for gloomy reading. We are by no means discouraged by this. Rather, we are focused on the long term. Dell'Oro's forecast for the global optical market between 2024 and 2028 is 6% higher than the previous five-year period. The analyst further states that the total router market will grow at a CAGR of 1% over the same period, while the enterprise portion is likely to grow at 2%. But what is more?
In the end, Ethernet switching for data center market, Dell'Oro predicts a CAGR of 7% in the five-year from 2024, with modest growth in the near term starting to accelerate in 2027. It seems that the industry overall is, to some extent, writing off 2024, but confidence in the future is really strong. The next wave is coming, and we believe that it will arrive just as we are closing our proposed deal with Infinera, of course, subject to relevant approvals. In short, the last fifteen months have been challenging, a picture that will remain valid for the rest of this year. However, our teams are focused on maintaining momentum, particularly in order intake for the rest of the year and on preparing for the future.
This is a transitionary year for network infrastructure, one in which we are taking strategic steps to ensure that we are well-placed to take advantage of the predicted upcoming investment wave. Thank you, and now I'd like to give the virtual floor to Sandy Motley, who Fixed Networks business. sandy, over to you.
Thank you. And hello to everyone, and thank you for the opportunity to share some of the specifics Fixed Networks business. we're here to talk about broadband for the future now, and this really means two things. First, we're expanding access and bringing broadband to more people than ever before. And secondly, we're not just delivering broadband, we're also introducing a future-proof technology that will last for decades and will meet the needs of generations Fixed Networks is in Nokia the undisputed leader in the fiber industry, and this leadership is a position that we've held for many years. First, our leadership is validated by industry analysts such as Dell'Oro, who's ranked us as the number one provider in XGS-PON market for five consecutive years.
Secondly, our extensive customer base is also a testament to our success. We serve over 400 customers globally, and that includes 50 of the world's largest cable operators. And now more than ever, we're also bringing in small alt nets and non-tier one operators across many countries. And these relationships are certainly an example of the trust and the value that we bring to the market. And finally, our technology and our offerings in this segment are not only cutting-edge, but they're also uniquely tailored to meet the specific needs and business cases of our customers. We provide the full portfolio of 10 Gb, 25 Gb , 50 Gb , and 100 Gb GPON technologies, so we have the right solutions for our customers that drive their operational efficiency as well as their business growth.
Before we dive into some of the market numbers, let's briefly cover how fiber to the home market is defined, some of the technologies, and the investment life cycle. For fiber to the home, PON is the most widely deployed technology, and that is passive Optical Networks. More than 95% of deployments around the world use that technology, and it's a point to multipoint technology from the customer's central office, where the fiber access node or the OLT resides, and the fiber begins its journey in this network. Typically, it's a one to 64 split, so 64 homes can be connected to that one fiber. Then the in-home product is connected to the network, and that is a fiber modem, which is the ONT, and that also provides Wi-Fi for distributing the signal throughout the home.
From a life cycle perspective, fiber will last for a 100 years or maybe more. Once you invest in fiber, the upgrades can become very simple. There's no equipment in the middle of the network, the equipment resides on both ends, so as the technology advances, you can upgrade the speed easily. Installing the fiber cables is the most capital-intensive phase of deployment. This requires significant upfront investment, and around 90% of the total cost is in this process. These costs include fiber optic cables, labor, permits, construction equipment, et cetera. The remaining 10% of the investment is in the active equipment, the OLTs and ONTs, and some smaller associated products. When it's time for an upgrade, the fiber remains, can remain untouched.
Essentially, you spend a similar amount of money then about every five to 10 years to upgrade as the technology advances. And as the market leader in fiber, we have a huge installed base. So if you look to the right side of the slide, we have deployed approximately 12 million OLT ports in the field, and remember, each one of these ports can connect to 64 homes. We began deploying with GPON back in 2007, and now the next wave of technology is starting to gain momentum, which is noted by the pink line. Today, we have about 3 million XGS-PON ports in the field, so there's still significant potential ahead of us for upgrading the rest of the installed base. You can also see that there's a third wave that's emerging on this chart.
Normally, it would have taken a bit longer to move to the next generation technology, but because we have invested in our chipset to combine both the 10 Gb and 25 Gb PON solutions in that single product, we already have 25 Gb PON-ready ports in the field. And operators are beginning to activate this 25 Gb PON service, but we are definitely in the early adoption phase. And there's one last point I'd like to highlight before we move off this chart, something that's not shown on the graph. Our installed base certainly can support much more than 12 million ports that we have shipped. And this is because our shelves and our cabinets are only about 60% full.
So over time, another growth area for us is we expect operators to completely fill their cabinets, and we will see card sales as a result. So there's still plenty of opportunity ahead, in terms of these upgrades. So on the next slide, I'd like to talk a little bit about where we are in the market in terms of deployments, and then also upgrades with those networks. So as of December 2023, 42% of the homes are yet to be passed with fiber infrastructure, and 71% of the homes have yet to be connected to that network. So this certainly highlights the significant potential for continued network investments. The chart in the middle highlights upgrade opportunities. In 2023, XGS-PON shipments represented only about 40% of Nokia's total PON shipments.
So again, growing adoption and advanced technology are happening and are real. And then if we look towards the right of the chart, among our installed base, about 12% of our customers are XGS-PON ready, with the infrastructure in place to support the technology, and then about 32% are already transitioning to XGS-PON. So this means that almost 45% of our installed base is either transitioning or is already using XGS-PON. So clearly, this demonstrates the upgrade is a real wave, is something that we're absolutely seeing, and it signals to all customers that they need... in order to remain competitive, they too will need to move to this, next upgrade technology. Nokia is the only vendor that supports all PON technology options.
We see that different operators have different needs, and therefore, we provide a full portfolio of solutions, unlike other vendors who might push a customer to one specific technology or another. XGS-PON and 25 Gb PON are available today and will satisfy the market for many years to come. But for services that require greater than 20 Gb per second, 50 Gb or even 100 Gb PON will be available. We're a key contributor to the standards, and focusing on 50 Gb , we have already brought to the market the first true 50 Gb and 100 Gb PON platform, our MF-14, and now we're trialing 50 Gb PON. The commercial product will be available the second half of 2025, where 100 Gb PON is currently in the technology proof of concept phase.
We're evaluating different approaches with Bell Labs as well as with our customers, and we expect to deploy that product around 2030. So it is important to have all of these technology options so that we can enable operators to compete better and to meet their business needs based on their own customer sets, but also based on their existing networks. Two recent announcements really illustrate this well. First, it's with Google Fiber. They are deploying 10 GB and 25 Gb, and they're evaluating 50 Gb. While Frontier plans to also deploy 10 Gb and 25 Gb but their next step will be to the 100 Gb PON technology. Our leadership in fiber not only is around opportunities for connecting homes and for upgrades, but also it leads to other adjacent opportunities and adjacent markets, and first I'll start with cable.
Our fiber solutions are being deployed with cable operators, and they're just beginning to a large extent their transitions. For example, in March, we announced with Service Electric, a leading operator in Pennsylvania, that they're using Nokia industry-leading broadband equipment to support a multi-gig service for their customers. More recently, Home+, a cable operator in Taiwan, is deploying our 25 Gb PON solution, making it the fastest fiber network in that region. Next on our journey of opportunities is rural broadband and government-funded programs, and this is probably the second major opportunity that we have, and it's around rural broadband, but also non-traditional operators, and partly driven by government funding.
The BEAD program, and I'll talk more about that in the next slide, is certainly critical for the U.S., but it's just one part of the broader rural broadband initiative. There's only about 50% of the U.S. homes are in urban areas, and then there's 50% that are a bit more challenging in order to provide the connections, and this market is evolving, with different investment models emerging, and smaller operators will indeed be supplying this rural service, and we see about 3,000 of those operators in the U.S., and as a result, you know, we have intensified our focus on this segment.
We certainly still support tier ones in a very significant way, but we also see that rural providers are important, and as a result, our primary go-to-market initiatives will be with our partners in support of this program, and our recent partnership with KGP is an example of our commitment to that, and our commitment to rural operators in North America. But there are also government opportunities and programs around the world. We have been successful with alternate network providers, AltNets, for example, in Europe, but also neutral hosts around the world, and these new kinds of operators certainly challenge the established players, delivering ultra-fast fiber broadband to the home, and our next journey on our fiber thread here is in-home connectivity, and the opportunities that this service provides.
So as operators increase their investment in multi-gig networks, it's really essential that that capacity is available throughout the home, and that's why we've invested in Wi-Fi solutions and operator-managed Wi-Fi. It enables operators to control, to optimize, and to troubleshoot their networks to ensure the best user experience at the lowest cost. And of course, enterprise opportunities and enterprise solutions are also key here. The same fiber solutions used in the home can also be applied in enterprise settings based on the capabilities, and the latency that exists now in these fiber networks, and this includes campuses and office buildings, and through something that we call optical LAN. This morning, we announced a partnership with Ruckus Networks, and we're offering an integrated fiber and Wi-Fi solution to deliver broadband within multiple dwelling units, offices, large venues, medical facilities, and other enterprise segments.
Fixed wireless access, as Federico had shared, also complements fiber in areas where fiber deployment just isn't feasible. But it also allows fiber operators to quickly provide broadband to their customers with this wireless solution as they start to build out their fiber networks, because that does take time. And then, of course, mobile operators can compete with fixed wireless access using their existing mobile networks to generate new revenues from fixed broadband. And of course, Nokia can offer both fiber as well as fixed wireless access solutions, again, a complete portfolio for our customers. And then the last segment, adjacent segment to speak to is the development of new business models. For example, our customer, Fibertime in South Africa, is successfully using fiber to connect townships.
They're using existing solutions in combination with a unique micro-payment model that's based on fiber minutes, and this was developed specifically for these kinds of informal economies, and townships. It proves that our fiber to the home solutions work even in low ARPU areas. There's many opportunities ahead for our fiber solutions, not only connectivity and not only upgrades. Let me jump and talk a little bit about the BEAD program. We've been hearing about it for over a year now, and now it's really finally becoming a reality. The program was launched in August of 2022, and the first year was really dedicated to the government and states identifying the eligible properties for this funding.
This was known as the Volume One phase, where each state was managing that individually with the government entities. Volume Two is the state-specific phase, and this is where the states are putting together their very specific plans with how they will administer the program and how they will allocate funds. Approval of these plans has been accelerating, where 39 out of the 56 states and territories have now received approval from the U.S. government to proceed. As this approval is put in place, that starts the clock ticking, where they have to have a final program in place in one year's time.
And the most exciting development is that, some of the states are actually moving into the grant window, and that means that they could allow applicants to apply for this BEAD funding, and Louisiana is the first to do so. And you could see on this chart on the right, the specific schedule and expectations, for Louisiana in this space. And we've already seen some potential applicants look to partner with Nokia, and we have over 150 companies that have expressed interest in working with us. So we expect to have a very busy time over the next months and years, related to the BEAD program. And the last slide that I'll speak to is, some recent public successes.
As I've mentioned, we've got a strong customer base with a lot of references across our portfolio, whether it be fiber, our access control or fixed wireless access. We have open access solutions, Wi-Fi, home controllers, and now, very clearly, digital campuses. We've been expanding in the cable area, as I mentioned, with Home+ and Service Electric. But, also in the U.S., we have a few other programs to announce. This morning, specifically, we have an exclusive multi-year market share agreement for fiber access with AT&T. This deal covers 100% of their OLT business through 2029, including the BEAD projects that AT&T will be involved with. And it's a commitment for them to also operationalize our MF-OLT platform with Altiplano. So we've locked in the entire market for several years.
Another significant win for us is Google Fiber, another key reference as well. We've announced a 50 Gb PON technology trial on top of our previously announced 25 Gb PON deployment. In Oceania, NBN has chosen Nokia to deploy its latest MF series optical line terminals with the Altiplano access controller, and this is across their NBN network. In India, we have business for well over a million units, each of fixed wireless access and our beacons, as well as 700,000 ONT units with a particular tier one operator. We have a 100% market share with them for fixed wireless access and for Wi-Fi 6, and about 50% of their ONT business.
In Singapore, Singtel has become the first fixed broadband service provider globally to implement the International Broadband Forum standard for Wi-Fi management, and that's the TR-369 technology. This adoption, along with our Corteca home controller, allows them to efficiently manage multi-vendor home broadband customer solutions and accelerate their rollout of Wi-Fi 10 Gb broadband services, and more. Over the past year, as I've shared, we've expanded our portfolio of fiber and in-home solutions, but we've also broadened our customer base. We've added 76 new logos, four neutral hosts, 13 MSOs, 17 CSPs, and 42 enterprise customers. So as we look ahead, the opportunities really are vast, and with our continued innovation and partnership, we certainly look to remain number one in the segment. We're not just building networks, we are a technology leader, and we are defining the access market.
With that, I will now hand it over to Vach Kompella, who leads the IP Networks Business Division. Thank you.
Thank you, Sandy. Hi, everyone, and thank you for attending this session. My name is Vach Kompella. I head the IP Networks Division in the Network Infrastructure Business group. I've been with the company since the startup days in two thousand and one, and I've been leading IP Networks for the past three and a half years. Let me start by giving you a brief statement about IP Networks. Our core business is delivering IP routing and switching products to service providers and to certain enterprises, enterprise verticals that require mission-critical systems. Think about it. Given our presence in all the top-tier service providers, in residential and business service routers, in mobile backhaul networks, and internet exchange carriers, we have estimated that close to every packet in the world passes through a Nokia router.
In 2023, we did about EUR 2.6 billion in revenue, coming off of record years in both 2021 and 2022. However, like for most in the industry, 2023 was a tough year. We see signs of improvement, which began in Q4 of 2023, as well as for year to date and full year of 2024, with orders and pipeline improvements. However, we believe that the real turnaround will come towards the end of 2024 and into 2025. We continue to lead the market for the fourth year in a row in edge routing. Edge routing is about 70% of the total routing business. As a quick refresher, edge routing is everything in routing except the core.
In other words, it's aggregation networks, mobile backhaul networks, residential and business service networks, internet gateways, and more recently, some classes of data center switches and gateways. We also maintain our number two position in overall routing, if you exclude the China market. We continue to maintain our focus in providing best-in-class service provider equipment as we grow market share in the CSP business. At the same time, we want to continue to expand our coverage of the core enterprise verticals we operate in. That is utilities, transportation, government agencies, and so on, while growing in new verticals, and we are positioned to target multiple use cases in the web-scale space, starting with hyperscalers, but looking at all the companies that offer infrastructure as a service. Our portfolio of products includes systems using our proprietary network processors, as well as merchant silicon from Broadcom and others.
Two distinctions that we hold that have helped us to rise to this leadership position have been our innovation and our quality. Today, our routers that use Nokia's FP5 network processor deliver the lowest power per Gigabit, highly sophisticated quality of service, and the highest scale and performance in the class of service provider routers. We are the first to support 800 Gb Ethernet in the industry, and the FP5 can even support 1.6 Tb per second whenever the optics are available. On top of that, we deliver line rate security that is quantum-safe. On the software front, we continue to make great progress in expanding SR Linux to support our ambitions in all domains.
It is a modern and open Network Operating System, or NOS, that has been designed for highly automated world of the future, with near real-time streaming telemetry, with APIs to access all provisioning and management functions. All the key applications are battle-tested, as they are derived from the rock-solid work that we have done with SR OS, the NOS that we have been delivering to our customers for the past 20 years, and we have just released EDA, or Event-Driven Automation, the most flexible automation platform for the data center and beyond, preparing us for scaling DC fabrics with zero error. EDA brings to networking what Kubernetes brings to cloud storage and compute management. As you all know, Kubernetes is the automation engine that powers GCP, Azure, AWS, and a whole host of other cloud environments. Coupled with these industry firsts is our enduring quality.
For 20 years, we've been selling into the service provider space and have never caused a massive outage. For example, like the one we saw last year in Canada or in Australia, and even the U.S. earlier this year. Our routers connect the majority of large power grid systems in North America and Europe, rail signaling networks, air traffic control networks, and so on. When I spoke to you last year, I emphasized that we're extremely focused on maintaining our lead in service providers while driving diversity in our customer base. On the CSP side, I said we would do at or slightly better than the market rate of growth with service providers. What is critical in this profitable but low-growth area is to capitalize on slips made by competitors to gain market share.
While I'm not at liberty yet to show you the customers we have won over the past four years, it's notable that many of these accounts were not Nokia house accounts prior to 2022. In other words, our product differentiation is amplified by either mistakes or strategic withdrawals by competitors as they try to navigate the service provider waters. Take the BNG, the Broadband Network Gateway. This is where your residential broadband service meets the internet. Nokia is arguably the leading vendor in this space, right when two phenomena are taking place. First, government-funded broadband programs like BEAD in the U.S. And secondly, the growth of fixed wireless access as a way to monetize 5G. Here, the security and sophisticated quality of service, coupled with quantum-safe networking, has sealed our position.
On top of that, we are the only vendor that can provide a single BNG to support both fixed access and fixed wireless access, lowering the cost of ownership dramatically. It is no wonder we are winning more broadband deals and see a bright future in the coming years. Or consider the internet exchange providers, who have traditionally made it their business to connect CSP networks. As the growth of subsea cable proliferates, the IXPs have found themselves in the enviable position of connecting these massive optical capacities from the landing stations to the terrestrial networks. They require a path to 800 Gb Ethernet, quantum-safe networks, and they would like to mitigate DDoS attacks on their networks. These are all key features of our routers and sadly lacking in competitors' products. The IXPs are a close-knit community, and when one finds good success, they share their findings.
The IXP community has been largely a competitor's stronghold, but I would like to say that we have cracked that wide open. We now have seven of the top ten IXPs in the world, and we continue to add to our customer base. I would like to reiterate that we see our strength in CSP space, that is continuing to drive our business. We will capitalize on our product differentiators and our competitors' missteps, and we recognize that geopolitics may be in our favor in some regions and against us in others. Again, we expect to grow at or a little faster than the market, while admitting that this segment is a flattish one. The second focal point for us was expanding our footprint in our traditional enterprise business, the OT or Operational Technology networks.
For example, the networks that utility companies build to connect their substations. I said last year that we would have a strong 10% growth in this segment. Prior to 2023, we had uneven success in utilities, transportation, and government agencies. So the first order of business was to ensure that we drove our success in these areas consistently across all geographies, and I'm happy to say we are succeeding. Again, reliability is a key factor, as well as diversity of interfaces that we support and longevity of products in these mission-critical, but very conservative industries. On top of that, we want to expand into more verticals that appreciate the robust, high-performance systems that we build. National Research and Education Networks, or NRENs, became a target. As you can see, we're beginning to make inroads in this space.
I'll comment on one of these, ESnet, the largest backbone network run by the U.S. Department of Energy, connecting all the major research facilities together. Not only did we win ESnet and replace the incumbent, we impressed them so much that they have volunteered to introduce us to other NRENs in the world. This is how we were able to get into and win CANARIE in Canada, GÉANT in Europe, and several others. We see a huge growth opportunity in this area that we did not play in prior to 2023. 800 Gb Ethernet, quantum-safe networking, and our reliability have been key differentiators in winning these deals. I continue to expect great things from the enterprise mission-critical verticals and look forward to continuing to post north of 10% growth as we drive new products into the market next year and expand into more verticals.
Finally, our third area of diversification is in the data center world. Last year, I said we would hyper grow in this space. Many of the deals we are winning here are a confirmation to us that we know what we're doing, even as we catch up with the impressive presence that our competition has established. As of this moment, we're deep in pilot trials with three hyperscalers and getting good recognition for our capabilities with others. We're working on our indirect go-to-market programs, and as Federico has mentioned, we have established partnerships with Kyndryl and Lenovo and are continuing to grow our global presence through partners. Various estimates put the TAM for this market at about 20 billion EUR, and it's certainly a great opportunity for expansion for our business.
But to tell you more about this exciting world of hyperscalers and data center operators, infracos and colos and enterprises that are building out their own private data centers, let me introduce you to Mike Bushong, VP of all things data center and IP networks, and formerly GM for data centers at Juniper. Mike?
Thanks, Vach. Hello, everyone. My name is Mike Bushong, and as Vach mentioned, I'm a data center guy. Most recently having been the GM, with Juniper's data center business. I want to spend the next few minutes building on the baseline that Vach set up. Let me start with just some data center basics. Vach talked a bit about the market, the takeaway there, big market, good growth, plenty of headroom to grow for Nokia. Now, if that's the starting point, the foundational question I want to answer today is pretty simple: Does Nokia have unusual permission to play in the data center market? Essentially, if we build it, will they come? Now, to answer that question, I want to frame this up around changes.
There's changes in the competitive arena, changes in how people evaluate and purchase networking solutions, changes in technology, and as I walk through this, keep one thing in mind: If even one of these conditions was true, that alone would be enough to create an insertion opportunity for Nokia. That all of these are happening at more or less the same time, that's why I'm bullish about our data center prospects. Okay, so, let's start with competitive perspective. To understand what's happening competitively, we got to start with customers. If you ask anyone who builds and operates a data center what they want, they're going to say three things: it just works, it's easy, it's affordable. That's it. That's actually what people want.
Now, what does it tell us about the current state of the data center switching market, that the number one criteria for purchase is it just works? It tells you that the current suppliers are struggling to do the basics. I want to show you a slide that one of our competitors has made famous. So this slide builds on an idea that Arista has used to litigate the case for product quality being the most important factor to consider in data center switching. This chart shows the total CVEs by major vendor. For the uninitiated, a CVE is a customer vulnerability exposure, basically a security issue that's reported according to NIST disclosure guidelines. Now, that's a government agency that's part of the U.S. Department of Commerce. Think of this chart as showing the number of customer vulnerabilities in the networking space by supplier.
Now, I want to be clear, CVEs are not a direct measure of software and hardware bugs. Not every bug, for instance, is a CVE. But if you were to review the actual list of CVEs that sits behind these numbers, you'd conclude that every CVE is indeed a bug. So while it's not a perfect measure of quality, it's a reasonable proxy for what's going on in our market. Okay, so what do we notice? The incumbents, they've lost the plot, and we aren't talking about some small difference in performance. Arista has a 13x advantage over Cisco and a whopping 20x advantage over Juniper. There's no surprise they're taking share in a market that just wants things to work. So where does Nokia stack up in all this?
We have our own 42x advantage over Cisco and a ridiculous 66x edge over Juniper. The punchline, and I want to be really explicit here, there's only two credible suppliers of quality data center solutions in the market. Now, is that enough to change how people are buying? Let's take a look at market share. So what does this chart show us? Our industry's incumbents lost almost 12 points of share since 2019. This is actually remarkable. In a market where things like cost are important, remember, it makes the top three list of things buyers care about, there's incredible incumbent advantage because the switching costs to swap out one supplier for another are typically enough to hold an account. And while it's natural to see some customer churn, what we're seeing in the data center switching space, this is like once in a generation.
That's before you consider the impacts of product life cycle. Think about this. Cisco, they're in the midst of a significant product transition from Nexus to Catalyst, from ACI to NDFC. As these products reach EOL, it creates a moment for customers to consider, "Do I stay with my incumbent supplier or do I look elsewhere?" Now, I've been on the road visiting customers globally for all, but literally two weeks since I joined mid-January this year. Our account teams, they walk me into all kinds of places, and yeah, I've walked into Cisco shops where the account manager basically briefs me with, "Guy with a Cisco tattoo, good luck." Shops that were 90% single vendor a few years ago are actively flipping the balance. It's like 90/10, becomes 70-30, becomes 50/50. Who's going to take advantage of that?
Now, since I just came over from Juniper, let me quickly cover what's going on there. The CVE chart speaks for itself, but obviously, they're working through the portfolio and channel conflict associated with the HPE acquisition, and regardless of what anyone thinks about their collective ability to execute, having to reconcile some of that in the core of their focus markets is creating uncertainty. We're already seeing channel partners lining up because they see the same thing we do, market share changing hands, and these conditions, they wait for no one, so they're actively bringing us deals because we're a simpler data center solution with way less uncertainty. Okay, so that leaves Arista. Let's be honest, right? They're the unquestionable market leader right now.
They've absolutely been on a tear with a combination of cloud titan success and incumbent fatigue, but the market fundamentally wants a strong second supplier for data center Ethernet. And look at how they've won, quality and software, the same things we do. If there's one thing the supply chain crunch did for everyone, it's put a little fear into sole sourcing everything. Now, as I mentioned earlier, changes like this in the competitive landscape are enough to create a meaningful growth vector. But all of this is happening at the same time that there's a wholesale shift in buying behavior. Look, it's impossible to ignore the fact that our industry is built on the backs of hardworking network engineers, and there was never a hotter recruitment period for network engineers than the late 1990s.
Those network engineers now, we're an aging lot, and as we move on for our network engineering careers, we need to develop some hypotheses about the bumper crop of next-generation networkers. Are they going to come out of university with networking degrees and vendor certifications, or are they more likely to be well-versed in all things cloud and familiar with an entirely different ecosystem of tools and techniques? Importantly, as this group takes their positions in the workforce, it changes the dominant user interface for our industry, and that, that's going to loosen what's been an incumbent stranglehold on the workforce... and we already see it. While Cisco's dominant in the enterprise, they're decidedly less so in the hyperscalers and SaaS companies that were born in the cloud. So how do we attack? If quality is the ticket to ride, operations is the reason to stay.
And in the management space, we've developed an operations platform that takes data center networking from configuration-oriented to model-driven. It's repeatable, it's fast, it doesn't require a certification to use. It's built around Kubernetes, so it's familiar to the next-gen networker, and it's multi-vendor, which gives us an insertion into non-Nokia environments. And by automating workflows, we create a bridge from legacy environments to cloud. We're more than a data center management tool. We're the operations boost for anyone who's tried and suffered for decades to automate. Okay, so we have changes in the vendor space and changes in the workforce, but you can't ignore the effects of technology. Look, AI is doing two things. It's driving a wave of architectural evolution that favors the new over the old, and it's diverting spend. So how do we participate?
If you thought network reliability was important for normal applications running on standard servers, imagine shelling out billions in GPUs and then watching it all sit idle because of a bug. "It just works" is even more important here, and frankly, if you can't make the standard stuff work, you're not gonna get permission to try with the heavy stuff. Our portfolio, it's based on Broadcom silicon. So as Ethernet picks up in these clusters, everyone's gonna want a couple of different paths to get there. We're in a good spot to carve out a second supplier position, and that's before you consider our pluggable optics, which make up, like, half of the total deal size in these large clusters. Having a complete solution, and that includes the WAN side, by the way, that gives us some really powerful levers to pull as we compete.
But I don't want people to think that AI is the only technology working its way through the data center. AIOps, Kubernetes, multi-cloud, these all have a role to play in our industry's ambition to be more automated. The operations platform I referenced earlier, that's the path. We've already integrated telemetry and workflow, and the migration to multi-domain, multi-data center, multi-cloud, it's already accounted for. Now, I just wrapped the technology discussion with a single sentence: Anything that resets the playing field is gonna bring a lot of opportunities into play. And speaking from a little bit of experience, I'd be shocked if new considerations are gonna yield old answers. Okay, so let's finish this up. Basically, should you believe everything I just said? Here's a few things to know. Vach already mentioned our traction in the hyperscalers.
That's a good place to start from, especially if there's a technology wave that's rolling through. We've got wins in Tier one SaaS, Tier 1 FinServ, top-tier healthcare, the top colocation companies. We've seen wins in GPU-as-a-service use cases. We're on pace to more than double our win total from last year already, and that's before you include increased go-to-market investment. We've been adding non-CSP salespeople with data center experience. That should only fuel the fire. We've already mentioned our progress with Kyndryl, Lenovo, and that's in addition to other ongoing engagements in the integrator space. I talked a bit about channel earlier. We've more than doubled our data center partners this year over last in just the first half of the year. Now, again, if even one of these industry changes was happening, I'd be bullish on our chances.
Having all three happen coincidentally, candidly, we couldn't fully predict how favorable the conditions have become for us. Incumbent fatigue was already setting in, which, frankly speaking, provided enough opportunity to give us a little institutional courage, but the share loss, the product EOLs, the quality problems, the supply shock, the HPE acquisition, the massive AI opportunity, it's just been one fortuitous turn after another. There have been two really great times to enter the networking market before now: just before the dot-com era and at the inception of cloud. When we look back a decade from now, we'll identify this period we're in now, the pre-AI era, as the third, and while a lot of the attention will indeed be on AI, let's not forget that there's a $20 billion market that's begging for a credible second supplier.
That gap is our opportunity, and if the traction over the last year is any indication, we're sitting on a winner. Vach, back to you.
Thank you, Mike, for sharing our data center vision. As you can see, we have exciting things planned for IP Networks in both our traditional CSP and enterprise spaces, as well as in the burgeoning new area of data centers. Now, without further ado, I hand the baton off to James Watt, who leads the Optical Networks Division. James?
Thanks, Vach. Good afternoon, everyone. My remarks today will first cover our organic Optical Networks business, and then I'll comment on the planned acquisition of Infinera. Let me start with an overview of the business. Last year, we had a record performance with almost EUR 12 billion in sales and a strong operating profit. As an aside, as is typical, this was a year later than what Vach mentioned for IP. We're the global leader in terms of the breadth of portfolio we have and our global reach with that portfolio. No other vendor matches Nokia and its ability to deliver Optical Network solutions around the world. To that point, six of the seven markets we serve worldwide are represented in our top 10 customers.
Speaking of customers, in terms of market share, we're number one in Europe, rest of APAC and India, and number two in MEA and Latin America. We're fast approaching a million coherent ports shipped lifetime to date and well over a thousand customers, of which almost a third use our switching platforms to aggregate traffic and/or deliver services.
... Despite all this, 2024 will not be another record year, given the industry-wide challenges that we've all seen. We are seeing increasing strength in orders, and that, taken with the trends mentioned by Sandy and Vach, and combined with the increased traction momentum, will derive from our improved differentiation. The trend will shift as we move through 2024 and 2025. You heard Federico comment on many of the emerging market trends. With Optical Networks being the foundation of all communications infrastructure, those trends translate into real-world needs for, and drive the evolution of Optical Networks. Our customers look to Nokia to enable scaling across the full range of their Optical Network applications.
Starting from their customer side, first, they see the need to transform their access or edge to support the move from 10 Gb to 100 Gb and beyond, using innovative, compact, low-cost, and low-power coherent pluggables, such as Nokia's 100 Gb ZR, to address the coming increase in residential broadband capacity mentioned earlier by Sandy. On the business services side, the continued shift to the cloud drives the need for significant capacity increases. Similarly, in the metro, as well as data center interconnect, customers are seeking to employ at least 400 Gb and even greater capacity if possible, with the demand for data center interconnect driven by the continued growth in cloud services, whether public, private, or otherwise. As that data moves into the core of their network, our customers are looking to deploy 800 Gb wavelengths across multiple thousands of kilometers to achieve the lowest costs.
It goes without saying that our customers' core networks connect with the subsea cables enveloping the globe. The need for bandwidth in the parts of the Optical Network is further amplified by the effects of AI, with optical technologies being the key way to deal with the combined needs of massive capacity, extremely low latency, and minimal power per bit moved. As we look at this picture across the entire network, it becomes clear that scaling needs to occur in two directions: ever greater capacities, wavelength speeds, and switch capacities for metro and core networks, and equally important, scaling down towards the network edge and access in the form of compact solutions such as Edge ROADMs and compact pluggable coherent transponders. Given those pressures, what we need to deliver is all that is required to bring cost-effective, high-bandwidth, high-flexibility connectivity throughout our customers' Optical Network.
Simultaneously, we need to ensure that designing, deploying, operating, and optimizing the network stays tractable for the customer, despite the size and scope of the resulting network. We call this scale made simple. While reconciling these two seemingly conflicting objective results in a long list of topics to be addressed, I want to cover the three primary topics here and make sure they're clear. One, we need the transponders to get the traffic on the network with the capacity needed. Two, the photonic line needs to get that traffic where it needs to be. And finally, the automation required to ensure operating the resulting network is tractable.
To deliver the capacity needed by our customers, we have and continue to deliver ever more advanced coherent engines, delivered in both embeddable and pluggable form factors, and hosted in modular platforms tailored to our customer operating environments, both telecom and data center. As a result, we have a portfolio of coherent pluggables, which our customers can also host on other platforms. We leverage our position next to the leading IP networking team in the world to deliver industry-leading IP optical management integration. Equally important for capacity, and essential to get the right bandwidth in the right places in the network, is the photonic line, which has and continues to be an area of strength for Nokia. Here, we deliver integrated C+ L-band systems with open control and spectrum management features at the high end, but also elements suitable for the metro, access, and edge.
To keep operations of these networks trustable, we deliver sophisticated software systems that leverage the latest software technologies, natural language processing, machine learning, and other techniques to provide abstractions that endure across the lifecycle, closed-loop automation, even if semi-automatic, and multi-vendor capabilities supporting the network service set and end customers of our customers. The majority of our investment is focused on innovating in these areas in order to create sustainable differentiation. Let's look at some examples. The PSE-6s is the most recent example of Nokia's ongoing commitment to, and execution on, technological innovation in coherent optics. Every DSP generation we've launched in some way brings a unique set of capabilities to our customer. We introduced the first single coherent 100 Gb coherent solution in the market, and follow that up with the first high-gain SDFEC solution.
Starting with our third generation of coherent solutions, we were the first to introduce a dual-line application-optimized family of coherent optics, with a super coherent line on one side focused on maximum capacity reach and the compact coherent optics optimized for pluggability and low power. We continued to introduce such innovations as probabilistic constellation shaping, and continued down this path. Our compact coherent engines combine our in-house vertically integrated silicon photonics optical front-end with a purpose-built DSP to deliver the lowest power consumption and offer the most cost-effective pluggable solutions for metro regional applications. Back to PSE-6s. With PSE-6s, we delivered the next milestone in super coherent optics, enabling scale of 1.2 Tb per second per wavelength, unmatched reach of over 2,500 km at 800 Gb speeds, and enables significant reductions in network level power consumption.
We've been shipping PSE-6s solutions since late last year. While the product teams did a fantastic job delivering the PSE-6s only two years after the PSE-5s, what has been even more gratifying is to see how our customers have made use of its capabilities. The public announcements you no doubt have seen in the recent quarters are testimony to the performance and promise this technology holds, and its ability to enable our customers to differentiate the service to others. Our first mover advantage here means we'll win more than our fair share of business in this highly competitive market. The PSE-6s has allowed us to return to our habit of setting world records with at least four current records around the world. These include, with Orange, we set the world record for subsea transmission of 800 Gb at over 6,600 kilometers.
With Zayo, again, 800 Gb over close to 1,900 km terrestrially. In addition, we achieved a world record with Turk Telekom International for 800 Gb over 2,300 km on a real production terrestrial network. And along with China Unicom, we announced another world record of 800 Gb over 3,000 km, this time using large effective area fiber. So 6s has us in the lead on the transponder front. One of our key investments on the simplification side of our strategy are the innovations underpinning our WaveSuite solution and how they help our customers optimize their networks, scale their networks, and monetize their networks. WaveSuite enables our customers to adapt the dynamic, unforecasted, and ever-growing demands of Optical Networks, given the drivers we saw before, with comprehensive solutions for network design, operations, and control.
It's a combination of an open, programmable management platform, which makes it easy to automate network operations, integrate with orchestrators, operations support systems, and business support systems. A reset of resource control capabilities that let our customers visualize network topologies for multi-vendor network integrations, and a comprehensive platform of ready-to-use applications that help our customers introduce focused solutions to optimize, scale, and monetize their optical transport. To show the value of WaveSuite, we worked with a research firm, Analysys Mason. They interviewed existing WaveSuite customers to find out how they performed their network management service delivery in the past, how that changed with the introduction of WaveSuite automation, and the economic impacts of those changes. The results were very compelling.
Economic highlights included 81% OpEx savings for service order orchestration, just over 50% OpEx savings for network management, and almost a third CapEx savings utilizing advanced network planning capabilities. And when we put this all together, what can we do? Tier 1 CSPs continue to be a very significant portion of our sales. We've been able to help several of these customers in recent years. Some of them are new to us, others have been with us for more than a decade or more. These customers form the foundation of our success in the market and respond strongly to the kind of innovative solutions we bring to market. For example, in India, our innovations have translated into more than 50% market share.
At Bharti Airtel, we're partnering with them to build a multi-terabit pan-India Optical Network to provide massive capacity to enterprises, operators, and hyperscalers, using our OTN technology for efficient aggregation, switching, and high resiliency. While growth in the tier one CSP segment is modest, we'll continue to leverage the combination of our product strengths and ability to execute anywhere to grow. In a related but different market, we've partnered with carrier neutrals in order to maximize their ability to monetize their network. They're often global providers of end-to-end digital infrastructure solutions for various operators, providers, and web scalers, leveraging their large fiber optic networks to provide open, transparent, and neutral access to cloud and internet services. Here, our ability to provide high capacity, low latency, and high resiliency solutions on one hand enable them to better monetize those on the other are keys to our success.
Security of communications is a hot topic, and an example of our capabilities in this area is on display at these service providers in South Korea and Singapore, as they leverage Nokia's quantum-safe networking technology to boost cyber defense for their enterprise customers. Earlier, Federico mentioned the proliferation of submarine cables. As we saw, in collaboration with OINIS, the international arm of Orange France, we set a world record on the Dunant cable. As an aside, the 800 Gb per second we achieved on that cable is eight billion times more capacity than the first transatlantic transmission by Marconi 122 years ago. Another example is the South Pacific cable, running along the west coast of South America. We are lighting this cable for Mexican service providers, the seventh largest mobile operator network in the world.
In fact, they purchased the very first pair, PSE-6s-based solution in all of Latin America, to provide capacity in this cable... as well as lighting the subsea cables. Where these cables land, we benefit from the explosive growth in the adjacent market by partnering with terrestrial service providers to build digital bridges, to straddle between subsea cables, or to bring access to web-scale platforms to local residents and enterprises. Overall, capacity and integration with the broader terrestrial network are key here. All those subsea cables and digital land bridges ultimately connect to data centers for cloud services. Our cloud customers are able to deliver fast, high-quality bandwidth services based on our solutions. The solutions they deploy ultimately enable them to supply co-location, hosting, and cloud services, including GPU-as-a-Service, and also allow them to transport AI traffic to and from training and inferencing workloads running on their platform.
Again, capacity and our ability to deliver them are key differentiators. Away from the public telecom space, we continue to grow our enterprise sales in multiple vertical segments, including energy, transportation, public sector, and research networks, just to name a few. We deliver these networks both by working directly with enterprise customers, but also work through system integrators, distributors, and channels, and our service provider customers who deliver private managed networks. Recently, this has been our fastest-growing segment, and we continue to see significant growth potential here. Our ability to deliver high reliability solutions across a wide range of platform capacities and quantum safe network are key elements of our success. A few examples: Terralpha, a subsidiary of SNCF. With them, we deployed an Optical Network connecting more than 150 points of presence over 20,000 kilometers of fiber.
Along the railway lines in France, they could offer Ethernet services up to 400 Gb. In Austria, with Illwerke vkw, we're lighting up Nokia's optical solutions along its high voltage transmission lines in order to provide 10Gb, 25 Gb, and 100 Gb mobile backhaul and fronthaul transport services. Meanwhile, in Netherlands, SURF is getting ready to connect Amsterdam and Geneva using our PSE-6s technology in order to accelerate the massive data exchange between the CERN particle accelerators there and the Dutch National Institute for Subatomic Physics. Our global reach, combined with increasingly strong portfolio, has us well-positioned to capitalize on the coming sources of traffic growth. I want to now spend a few minutes reiterating the thinking behind the planned acquisition of Infinera. At a top level, the key drivers are threefold: scale, footprint, and technology depth. First, scale.
As you will have gathered from the comments above in our organic business, the world of Optical Networking is technology-intensive and requires continuous investments across a wide range of technologies. With the software and silicon, we see broadly across telecom, but also are more specialized topics like silicon photonics and even more specialized materials. On one hand, as we just saw, these technologies can be deployed in many markets. The investment required to keep up the demands of our customers, and to do so in a way that is leading edge, despite the large competitors in the market, is still significant, and this is only exacerbated by the ever-increasing pressures on the Optical Network. The scale is essential to both deliver what the customers need, when they need it, and be able to generate a reasonable return on the investments made.
If you then take our combined scale and put that against the optical market forecast, Nokia is well-positioned to participate strongly in this growing market. Second, footprint. Two businesses, highly complementary geographic footprints, as Federico mentioned. Nokia's strength in optical has historically been in Europe, Latin America, MEA, and Asia Pacific, as you can see on this slide. Roughly 70% of our sales are in these areas. The U.S., while an area of focus, was still less than 20% of sales last year. Infinera, on the other hand, has a strong position in the web-scale market and the U.S. market, with this region making up over 60% of its sales. The combination significantly enhances our presence in the optical market across regions and globally, and noticeably broadens our exposure to the fast-growing web-scale market.
And finally, from my side, I'm really excited about the combination we're able to do for our customers. The first thing I should say, and this is in the DNA of Nokia, as we progress through this acquisition and the integration processes, we will look after customers, ensure they see as little disruption as possible. We've done it before in optical, when we brought Alcatel and Lucent together, and we are committed to doing it again this time. The combined business will have a very special mix of capabilities to deliver the best possible outcomes for our customers. A few examples. We'll have an expanded DSP team that will really help us accelerate our product roadmap.
Separately, we've both been able to improve our competitive position in recent years, but together, this means we'll be able to make fewer trade-offs when allocating resources across competing development efforts and deliver broader and more complete product progress to our customers. This will be supported by our fundamental research within Bell Labs, ensuring we have access to real cutting-edge innovations in the optical and material sciences domains. The combined team will also have a more complete technology capability across different material sciences, used in optical, and allowing greater vertical integration and an increase in our capabilities and pluggables, and the offering towards inter-data center and intra-data center applications. Finally, the integration capabilities will enhance our manufacturing capabilities with Infinera's U.S.-based fabrication and packaging expertise.
Taken together, scale, footprint, and increased technological depth will allow the combined organization to deliver greater innovation, accelerate our product roadmap, deliver customers a broader and more complete offering that would today that we can deliver globally, and to do all that while generating respectable financial return, and with that, I'll hand it back to you, Federico.
Thank you to Sandy, Vach, Mike, and James for their presentations and their dedication to driving success for NI. Before we turn to the Q&A, let me just conclude with some of the points we hope you can take away from Fixed Networks, we remain the market leader, and we see great opportunities ahead with a strong demand for fiber, government funding programs ramping up, and operators upgrading to XGS and 25 Gb port. In IP networks, we have demonstrated over the past two decades the quality of our products in CSP edge routing, and our investment has continued. Indeed, we even see opportunity to gain share. We are also building on this with good momentum in enterprise and web scale, something we are particularly excited about.
Just in case you missed it, there are opportunities in data center for our IP networks, as well as Optical Network business. Finally, in Optical Networks, with the acquisition of Infinera, we believe we will be in the right place at the right time to increase our scale, accelerate our product roadmaps, and strengthen our position for future growth, diversity in web-scale, and artificial intelligence investment. By combining all these elements together, I believe NI has the potential to deliver mid-single-digit growth and improve our operating margin to a mid- to high-teens level in the long term. Thank you very much. So, David, let's start the Q&A.
Thank you, Federico, and thank you to the broader team for all your presentations. We will now move to the Q&A session. As a reminder, these events are intended to be focused on our products, technology, and strategy rather than a financial update, and I would ask that you focus your questions towards our network infrastructure business. I would also just highlight that those that are joined onto the webcast, you need to join the phone line, the conference call dial-in, if you would like to ask a question. Operator, could you please give the instructions for the Q&A session?
Webcast, please remember to mute the audio on your computer before asking your question, as there is a 30-second delay. To ask a question, you may press star then one on your telephone keypad. If you are using a speakerphone, please pick up your handset before pressing the keys. To withdraw your question, please press star then two. I will now hand the call back to Mr. David Mulholland.
Thanks, Drew. We'll take our first question today from Jakob Bluestone from BNP. Jakob, please go ahead.
Hi, thank you for taking the question, and thank you for the presentations as well. I had two questions, please. Firstly, on Mike's presentation around data centers. You talked a lot about some of the external factors that are providing an opportunity for Nokia to take more share. I'd be interested in hearing what's changing in terms of what's changing internally. So are you making changes to your products? And I guess as a sort of fresh pair of eyes, what do you think was holding Nokia back in the past, and what's changed to enable it in terms of its products to take more share down the road? And then just secondly, maybe if we can just have an update on-
Sure
... the Infinera process itself. Can you maybe give an update on the regulatory side? Anything you can share, particularly around perhaps antitrust or any other issues, so far? Thank you.
Okay, Federico, maybe you can start?
Yeah. I will start with the second, which is easier. And then on the first, I will give my view, and then I will pass to Vach or Mike. So Infinera, we are going through the process of regulatory approvals. We are following all the steps that are needed. As we announced, some months back, this is going to take somewhere between nine to 12 months, and we expect the closure by first half next year. Nothing new to report. I mean, we're going through the process, without any... nothing to report. On the data center, it's not that something has been holding us back. No. We were the new entrants some two, three, four years ago, and we started to enter into the data center market base, banking on quality on our product, our technology.
We started to win, but in this market, one of the important things to realize is that especially when you are a new entrant, you have to develop from scratch, and you have to develop tailor to your customers in the hyperscaler space. Now, we have been gaining momentum, and we feel really confident that we have a very good product from a quality and technical viewpoint with SR Linux and on the management system. But I will let Vach and Mike elaborate a little bit more.
Thanks, Federico. Yes.
Thanks, basically.
Just a follow-up on what Federico said. We have been in this space. We started developing SR Linux about five years ago, so it's not something new in terms of you know the process of engaging with data center opportunities. It's just that you know the set of features that are required in the data center are somewhat different, and they evolve much faster. So the question of how we prepare ourselves for the data center space is really to build the right type of infrastructure, the right type of products.
... and I believe that SR Linux serves us in good stead because of its much more flexible and open architecture because of you know some of the recent progress we've made in data center fabrics and a much more flexible management platform for data centers with EDA. I believe that these are all things that are helping us to gain share. The process of entering a new market also takes time because you have to not only insert yourself at the right point, as Mike was talking about, there are opportunities for insertion now, but also in terms of familiarity of customers with our products. They are...
You know, if you talk to service providers, they're very well-versed in, you know, our product portfolio. But when you talk to enterprises, on the IT side, on the data center side, not quite so familiar. So we're building, you know, more credibility in that space, engaging partners. So there's a whole process that goes on behind this, and I think we've shared, you know, what we're doing in this space.
Thanks, Federico Guillén and Jakob for the question. We'll take our next question from Simon Leopold, from Raymond James. Simon, please go ahead.
Thank you very much for taking the question and for doing this call today. I wanted to see if we could unpack a little bit on this concept that Federico talked about, about Clean Networks and sort of the geopolitical aspects. It seems as if a lot of the press coverage focuses on mobile, and given today we're focused on network infrastructure, I wanted to see if you could step back and help us understand how you're sizing these particular geopolitical opportunities for network infrastructure and where we stand today. Thank you.
Okay. So we are technology people, we are not politicians. We are here to serve our customers, but I mean, reality is that the customers in certain parts of the world have some restrictions to some suppliers, and that's why they create what they call Clean Networks. Imagine a particular customer in a region somewhere in the Middle East that wants to do business with a hyperscaler, and the hyperscaler want to make sure that his traffic is going through a particular set of vendors or excluding a particular set of vendors. That's what we call Clean Networks. And that's where, as we have the footprint we have worldwide, in many cases, what happened is that service provider is choosing Nokia in order to build a network that is then providing the service to the hyperscaler.
That's, that's what it is. And that could be optical IP, mostly, because of the big traffic that they're carrying.
Did you have a follow-up, Simon?
This opportunity has really manifested itself in its business. Because you've talked about being able to outgrow markets, and I presume that these opportunities are a key element of that and so I'm trying to get a little bit more quantification as to where this stands and how you're sizing the opportunity.
Yeah. What we're seeing is that the world is evolving into traffic that is where data center connectivity or data center interconnect is having more and more weight. The traffic that is generated little by little is getting momentum and having a significant weight on the total global traffic. As I tried to explain in one of my charts, to reach at the moment in time, it's going to reach something like 25% of the total traffic worldwide. Well, that is precisely what happens, that if you consider that traffic and you have to build new networks to deliver that traffic, and part of it is trusted into certain suppliers that have less restrictions than others, and we have the right level of people across the world.
On that, especially in optical, I can say that we have teams in all the regions, and therefore, our position in APAC and in EMEA is better than my dear American competitors. We have an opportunity that is well quite sizable for us.
Thank you, Federico and Simon, for the question. We'll take the next question from Sebastien Stawowycz from Kepler Cheuvreux. Sebastien, please go ahead.
Yeah. Hi, everyone, and thanks for taking my question. So regarding the strategic deal and this morning with AT&T on fiber access, could you please help us understand the potential size of this project for Nokia? Is this a significant expansion to your relationship with AT&T, and when do you expect the deployment to start to kick in, is it the end of this year or 2025 ? And the second question is on the core router market. You have been looking at this market for some time, but your market share remain at a relatively low level today. So how should we think about your positioning and the dynamic in the core router business for you? Thank you.
What do you mean low level?
The-
So I will let Sandy comment.
Yeah. Thank you. The call-
Go ahead, Sebastien.
Below your market share in edge.
Ah, in core router.
Yeah. Thank you.
On AT&T, well, basically what we have announced is one more step in our long-term relation with AT&T. We have been providing and being sole supplier of AT&T in access for many years, since the copper times, and then evolving in different technologies. From ADSL, we went to ADSL2, to ADSL Plus, to VDSL, to vectoring, to G.fast, to GPON, and now to XGS-PON. We continue the journey with AT&T, with the successful new technology introduction. I will let Sandy comment a little bit more. In terms of core and edge, well, basically, yeah, I mean, as Vach explained, 70% of the market in IP is on the edge.
That's where we have most of the focus, but of course, with FP5, we're better placed now to compete a little bit better in the core, but again, these things take time. So, Sandy, do you want to comment on AT&T, and then Vach on the core question?
Sure. So, with AT&T, as Federico said, we are continuing our strong relationship with them and continuing the exclusivity that we provide to them in terms of our OLT product. So this really isn't a change, this is just a continuation through 2029. And, you know, we're thrilled to be able to announce that we're keeping the competition out of AT&T's network. And, again, we continue to have the strong relationship and the strong partnership with AT&T, despite, you know, some other changes that have gone on in their other networks.
Did you have anything to add on the core router piece, Vach?
Yeah, sure. So, you know, with the FP4 and FP5 series of routers, we have extremely good performance, a lot of capabilities that really cater to the edge of the network. But with the higher speeds that we are now able to deliver with FP5 especially, you know, it turns into a very meaningful core router. One of the challenges, though, is that FP5 is designed for really high performance, so deep buffers, extremely good for a well-designed core. But not everyone is looking for a, you know, high-performance core router with deep buffers. They're looking to manage their CapEx by spending a little less on the core, where they just want speed, and they don't worry too much about buffers.
I would say that's maybe a bit of short-term thinking in terms of, "Let me just try and build a core without, you know, adequate buffering." But, you know, we understand that there are financial reasons why some operators will choose to cut costs on the core routers. We have plans for a low-cost core router that will be coming out next year, and it'll help us to address more of the core router market beyond, you know, what we can address with the FP5. I'm looking to see a bigger uptick in our core presence.
Thanks, Vach and Sandy, and thanks, Sebastien, for the questions. We'll take our next question from Felix Henriksson, from Nordea. Felix, please go ahead.
Hi. Thanks for taking my question, and thanks for the presentation. I wanted to continue on the data center and web scale topic, and hear your thoughts about where exactly we're standing in the investment cycle when it comes to both data center interconnect and intra data center optics. Should we already expect acceleration in demand already in 2025 ? And how do you see yourself positioned to capture market share, specifically when it comes to the transition from 400 Gb to 800 Gb when it comes to data center optics and switches? Thank you.
Wanna start, Federico, and then-
Yeah, I mean, I believe we can go directly to Vach and Mike. Yeah, on 400 Gb , 800 Gb, we were the first one having a router with applicable on 800 Gb, and we are really confident that when the traffic start to grow in the data center, we're going to be very well placed on that. But Vach and Mike, please help yourself.
Sure, I'll, I'll start. First, you know, with the portfolio that we have, you know, we're well-positioned to start with data center gateways and data center interconnect. We've seen quite a bit of success in this space, 'cause not only are they looking for higher capacities, but they're also looking for quantum-safe networking to connect their data centers together in a secure manner. Throw DDoS mitigation on top of that, and we have a great solution for the data center gateway and the data center interconnect.
I think on the optics, the way that data centers sort of move is they move their entire data center from, you know, 10 Gb to 100 Gb , 400 Gb and now they're looking at that next wave. As capacities on the data center side increase, especially with AI traffic, that will be the next phase. But I believe we're getting ahead of ourselves a little bit, because today, most of the large hyperscale sort of environments are going to 400 Gb . And I think 800 Gb is down the pipe, but... And we're preparing for it, but it's maybe a little early for that.
And then some of the next tier infrastructure companies are looking at 100 Gb, multiple 100 Gb with 400 Gb breakouts and so on. You know, Mike, if you wanna jump in, and then I think James, you probably should talk about the optics themselves.
Sure. So NVIDIA will drive the 800 Gb with CX-8, so there's a bit of a dependency there on what's happening on the NICs side. You'll see 800 Gb adoption, primarily in AI clusters initially. How far that moves down market will depend heavily on sort of what happens with digesting 400 Gb in non-hyperscale type accounts. But as Vach says, I think that there's still some time before that actually plays out. Right now, people are still working through 400 Gb and dense 400 Gb on the data center switching side, and then there's implications there on DCI.
James, did you want to add on the optics side?
Yeah, I was just gonna say this. I agree with Mike. AI, the AI infrastructure is what's putting the most pressure on the intra DC optics. And then the second point I make is that's something we'll come back on when we get the Infinera deal closed.
Thanks, James, and thanks, Felix, for the question. We'll take our next question from Tim Savageaux, from Northland Capital Markets. Tim, please go ahead.
Thanks, and good morning from the West Coast. My question's on the optical side, and it kind of follows the recent comments there. If we look at some recent market developments at Lumen, for example, and building a big, maybe very big, data center network for Microsoft and others, you know, how does Nokia assess that opportunity, especially given the Infinera deal, where there's quite a substantial installed base there? And do you see other opportunities like that for optical transport between data centers? Maybe not necessarily with the cloud guys themselves, but with their network partners. Thanks.
Hmm. I assume we'll maybe direct that straight to James.
James can start, and then I will probably add something.
Okay.
Obviously, we're not in a position to comment more directly on Infinera's business today.
Yeah, of course.
But yeah, high level points.
I won't comment on the Infinera, that specific deal, but I think the general topic of service providers building infrastructure dedicated, if you will, to a customer or even a customer set, is a significant opportunity, and one we've been taking advantage of. Sometimes you'll hear this called managed optical fiber networks. And again, there you know, if you think outside of North America, there are many places where the web-scale players can't operate networks, and they partner with a service provider to do it. As Federico mentioned earlier, we're well-positioned to capture those, given that, you know, that those are our customers in general that they're working with, and we've done that with most of the web-scale players in one place or another.
It's definitely an opportunity, and definitely one that we can capture a good chunk of.
To complement what James just said, as I stated during my presentation, the combination of Nokia Optical plus Infinera plus Nokia IP Data Center is going to place us in a fantastic place in the right time to have the portfolio that this growth is going to demand. So we are really excited about that combination, and the fact that, yeah, we're seeing a softening of the market in 2024, but we're seeing a wave coming in front of us that is going to be starting to grow right at the time in which we're going to have a very much stronger portfolio.
Thanks, Federico and James. Tim, did you have a follow-up?
I did. I think there was some mention of being fairly far along in trials. I don't know if it was three hyperscalers. Is that... I wonder if you could talk a little bit more about that. I know you had a pretty sizable data center switch that you're expecting to ramp, I don't know, kind of around now. But maybe relative to that, or what sort of opportunities are you pursuing there? If you can, you know, size them any way and talk more specifics about timing, that'd be great. Thanks.
Vach, do you wanna-
Sure.
Start on that one?
Yeah. The reason we didn't mention names is because we can't mention names. You know, with the hyperscalers, nothing is quick. They run extremely large networks, and they're very sensitive to, you know, the quality and reliability of the systems they deploy because of how much business is carried over their networks. So, all I can say is, pilot programs take a very long time. Their feature lists are fairly extensive, and they take a long time, and all the testing that goes in, in order to run the enormous networks they run, with quality. So, I can't say anything more about the specific hyperscalers.
But I could say that, you know, we're well on our way in these pilot programs that gives us the confidence that we are building the right products. We have the level of feature development, velocity, the quality that is giving these hyperscalers confidence to use our products. Yeah.
Thanks, Tim. We'll take the next question from Sandeep Deshpande from JP Morgan. Sandeep, please go ahead.
Yeah, hi. Thanks for letting me on. Mike, I have two questions. I mean, I think both have been addressed before, but I want to raise one point. In terms of these Clean Networks that you talk about, in Europe or in countries where the Chinese equipment suppliers have been banned, is there considerable share to be taken in routing or in optical that still exists with those players, and thus there is a real opportunity in the market there? That's my first question. And how much is that opportunity? And then my second question is on the routing side, actually. I mean, Nokia broke into routing quite a long time ago, in the data center market, quite a long time ago, as such, really. Why has it taken...
Has it taken the time that Nokia expected to improve its market share there, or is it taking longer than Nokia expected? Thank you.
Yeah, starting again with the second. Yeah, we knew from the beginning that the entry into the data center space was going to take time, was going to take resources, and we were ready to put the money to invest on it. And we're very happy because the product that we are now having has the right level of quality, and it is starting to have the features, and it has certainly the flexibility with SR Linux, and all our customers are delighted. I have talked to most of them in the last couple of months, you can imagine, because of the other topic of Infinera, and they all mentioned about the, how happy they were with the trials that they are doing with our data center solutions. On the other topic, that was, I guess-
The Huawei.
Oh, yeah. Sorry, the Clean Networks. Yes. On the other topic, how big is it? I cannot tell you. Obviously, what happened is that in routing and optical, Huawei in Europe, which is where your question was referring to, didn't have the market share they have in other parts of the world, of course. So the opportunity has a size that is less than that one could think, especially in IP and optical. Because obviously, those are areas, especially IP, more sensitive than others when they started with banning Huawei. It takes time, and then there is a fair competition between all the different suppliers, and we are gaining in the CSP space.
So what I expect, and this is a target to batch on the team, is that we get more than our fair share of that type of replacement, and we are seeing good momentum there.
Thanks, Sandeep. We'll take our next question from Richard Kramer from Arete. Richard, please go ahead.
Thanks very much, guys. Federico, I have three quick ones in succession. Let each of you answer. Federico, your team mentioned a range of customers, but can you tell us the portion of NI business which is addressing mobile operators? Under, you know, the prior CEO, under Rajeev, we heard a lot about cross-selling, but it was never clear whether selling edge routing or optical to mobile operators was a combined or a distinct sale. One maybe for Sandy, is fixed wireless access still growing in the U.S., or do you consider the largest end markets to be now outside the U.S? And maybe you can talk a little bit about how you'd address BEAD spending when a lot of it comes from very small customers. Is that gonna be directly addressed or through channel partners?
And then lastly, on the hyperscalers, maybe for Federico or others, are you confident in being able to announce multi-product deals with these firms that are testing your products? And is that in your growth and margin plans, and do you think they're going to be willing to adopt your proprietary silicon, or are they instead looking for specific piece parts from Nokia? Thanks.
I'll start on the-
Yeah, I will start on the mobile, operator. So, what is a mobile operator? I guess you're referring to mobile-only operator. There are very few of those in the world. And most of the operators are fixed mobile converged operators, and we are supplier to all of them. The thing is that, we have a... I mean, most of the meetings I have with customers, that come to Helsinki or where we go to meet them wherever in the world, I go together with, mobile networks, in many occasions, a certain level. Then, the purchasing separates, split the two. It's rare to see a deal where they having the same RFQ, mobile and fixed, standard fixed, Fixed Networks, but the customer is the same.
So CTO, CEO, they all work with us at all levels. Or even on mobile-only operators, of course, we sell mobile backhaul, we sell the MEC, we sell even fixed wireless access. So we have a relation with the same customers that mobile networks have, obviously. So the second question was for Sandy?
Yes.
Okay, so in terms of fixed wireless access, we do continue to see growth in the U.S. in this space. And it very much is tied to what you were just talking about in terms of the partnership with mobile. Because some operators, some large operators, you know, have very significant business, millions of subscribers in this space in the U.S., and they're running out of spectrum. So millimeter wave will be key for them. And we work with the Mobile Networks team on optimizing features, capabilities between their radios as well as our fixed wireless access in-home product. So, you know, very much we see this as a growing business. You know, CAGR is very significant for FWA in the U.S., but as you mentioned, also outside of the U.S.
You know, with the India wind that I spoke about, and in other parts of the world as well. So, it's a significant part of the growth scenario for broadband. And then you had a question about BEAD, and whether this... You know, because it's a lot of small operators that indeed are, you know, focused on deploying in rural areas. Indeed, we are working with many partners. The solution for rural is really quite different than that for tier one. It takes a number of companies together to be able to deploy these networks and to be able to work, whether it's operating systems, whether it's services, deployment. And we are very much engaged with a lot of different players in that scenario.
You know, we're partnering with, like, 150 different companies on plans for deployment of BEAD in the U.S., in the rural segment. Our partners will be a big part of the play there. We're certainly strengthening those relationships.
Yeah, yeah.
Another hyperscaler question?
I'm confident to have multi-product deals with hyperscalers in the future. Don't tell my boss, but yes, I'm really confident about that. That's the reason why we are doing what we are doing. I already explained that we are getting there with a fantastic data center portfolio. We have data center interconnect capabilities in James' portfolio, and on top of that, we have the portfolio of Infinera coming. The relationship they have with the web-scalers in parallel to ours, provided for, of course, regulatory approval. I mean, this is a perfect cocktail to be a much better player in that space. Yes, I'm looking forward to it.
Thanks, Richard. And we'll take our next question from Sami Sarkamies, from Danske Bank. Sami, please go ahead.
Hi. In the presentations, you were referring to starting market recovery in late 2024 or 2025, based on orders and intake that has been growing since late last year. Can you name some of the main near-term pockets of growth in terms of regions, products, or customers, please?
I mean, I cannot name my customer. I can tell you that what happen normally is that revenues follow orders, and we're starting to see three quarters in a row, we are having a book-to-bill that is more than one. What happen normally is that when we see a peak of orders in the fixed space, that is a symptom of a later recovery, one or two quarters later in IP and optical. Why? Because if you build up more fixed network, then the traffic grows, and then normally what you need, depending on how the network was configured or designed, normally another wave of investment in capacity comes.
On top, what you have as another trend, is the demand of traffic by the AI data centers, and this is something that goes in parallel to the growth of fixed. So all those things, without being able to tell you exactly when this is going to start ramping up, make us think that it's going to be somewhere between the end of this year and first half of next year. We see all the elements for that to happen again.
Did you have a follow-up, Sami?
Yeah, maybe a question for you, David, actually. You spent a lot of time regarding the data center opportunity. Should we expect that you will am- ... and break kind of the enterprise sales into sort of real enterprise sales and then web scale sales, so that we can track the progress?
Like, you'll have seen in our Q4 presentation, we already started giving you some context for what the makeup of enterprise is, with the chart that we gave you. Whether it becomes a metric that makes sense to track quarterly will depend on the scale, because it can be very lumpy within the different pieces, quarter- to-q uarter. But over time, certainly our ambition is to give you visibility. And as I said, that's why we already started doing that, with giving you the breakdown of what that enterprise sales was in 2023. So it's certainly on our discussion as to how can we help you understand where we're going and what's driving it.
Thanks, Sami. We'll take our last question from Simon Leopold, from Raymond James. Simon, please go ahead.
Great. Thank you for letting me back in the queue. I think these are hopefully easy ones. One, I just wanted to clarify on the data center and hyperscale opportunity. I presume we're talking about front-end applications and not back-end opportunities of AI clusters. That's the first one. The second one-
... is, I'm not as familiar with this, the CVE that Mike presented, but wondering if that's statistically meaningful for Nokia, given it's a relatively smaller sample size than Arista, Cisco, and Juniper. So if you can help us understand how to interpret that data a little bit more. Thank you.
Vach, do you want to start, and maybe even James comment as well, given some of the optical areas?
Mike.
With the second element.
Yep. Okay, Vach.
What was the first question again, Simon?
It was the answer on opportunities. You talked about the AI cluster back end or the data center front end, in terms of the use cases.
I believe we're really going after both, right? The question of whether we can penetrate the AI data center back end is really an evolution of how that back end transitions from InfiniBand, where NVIDIA has a fairly solid lock, to Ethernet-based solutions. Where, again, NVIDIA has their Spectrum-X class of switches, but there's increasing interest to have multi-vendor solutions in that space, a more open standard. And the Ultra Ethernet Consortium is, you know, really developing Ethernet standards for managing AI workloads. And it has everything to do with performance and how to deliver AI workloads without traffic loss. So really geared towards the specific problem of: how do I carry AI workloads securely?
Securely, I mean by without packet loss over an Ethernet fabric. This is, I think, a very interesting development that will open up the back end, which currently is very much InfiniBand, but also InfiniBand has its limitations in terms of speed and diameter of the network. So we're looking to play in both spaces. There are, of course, AI opportunities that we have invested in, as Mike mentioned, where the GPUs service wins that we have in which we're in the data center back end also. But the immediate sort of opening for many vendors in this space is, of course, the data center gateways that interconnect data centers.
Thanks, Vach. And, Mike, do you want to address the CVE question?
Simon, so the basic question is, is Nokia's footprint related to the overall size of the CVE differential? The short answer is, it's probably related, but SR OS has a one or two share outside of China, and so you see good numbers there, and that looks pretty apples to apples with, like, a Juniper or Cisco in the same space. And so what you see there is, like, orders of magnitude difference. Do I think a 66x advantage over a competitor holds? You know, should that really be a 42x advantage or a 35x advantage?
I think you could probably argue the numbers, you know, could come down a little bit over, depending on how you look at how they're deployed, but you're still looking at, you know, 20x, 25x, 30x, 35X difference. So think of it more as, like, the directional. And if you were to look at just overall experience, the actual turnover of incumbent suppliers, and the market share numbers kind of play this out, suggest there's something that's going on beyond there. So I use it as a method of demonstrating that there is a thing there. There is a quantifiable proxy for that thing.
And then, you know, certainly as we expand our footprint, what we expect is that SR Linux, while more nascent than SR OS, will track with SR OS because we're leveraging our investments on the routing side to deliver the switches required to go after the DCI side.
Thanks, Mike. And thank you to all of the NI management team for joining us today, presenting and answering the questions. Ladies and gentlemen, that does conclude today's event. I would like to remind you that during the event today, we have made a number of forward-looking statements that involve risks and uncertainties. Actual results may therefore differ materially from the results currently expected. Factors that could cause such differences can be both external as well as internal operating factors. We have identified such risks in the Risk Factor section of our annual report on Form 20-F, which is available on our investor relations website. Thank you all for joining us.