Good afternoon everyone. If you'd please find your seat, we'll be starting momentarily. Also, as a courtesy to our presenters, why don't you silence all your electronics at this point. Thank you.
All right. Good afternoon. Welcome. For those of you who don't know me, I'm Paul Glaser. I'm the Head of Investor Relations. Welcome to HPE 2025 Security Analyst Meeting. Thank you for joining us here in New York and thank you to those. Who are listening on the webcast. In our executive session today, you will hear from Antonio Neri, President and CEO, and Rami Rahim, EVP, President, and General Manager of Networking, and Marie Myers, EVP and Chief Financial Officer. Following the presentation, we will take questions. From the live audience moderated by Shannon Cross, our Chief Strategy Officer. Before I pass it to Antonio, let me start with the disclosures. This event may include forward-looking statements involving risks, uncertainties, estimates, and assumptions. If the risks and uncertainties ever materialize and the estimates or assumptions prove incorrect, our results may differ, perhaps materially, from those expressed or implied by such forward-looking statements. HP assumes no obligation to update such statements. Please find more information regarding our forward-looking statements on our website at investor.hpe.com. This presentation also includes presentation of certain non-GAAP financial information. While we believe that providing these non. GAAP financial measures alongside the corresponding GAAP measures provides a supplemental view to understand our historical and prospective operating performance. These non-GAAP financial measures may have limitations as analytical tools. Please note that these measures should not be considered in isolation or as a substitute for analysis of HPE's results as reported under GAAP. With that, let me turn it to Antonio.
Good afternoon and welcome. Thank you for joining us today. Whether you're here at the New York Stock Exchange or watching us on our webcast this morning, I joined my colleagues to ring the opening bell to celebrate the 10th anniversary of Hewlett Packard Enterprise, which we mark on November 1st. Throughout the past decade, the IT industry has experienced dramatic change driven by the on-ramp to digital, the growth in cloud, and the explosion in AI.
I am personally incredibly proud of how we have transformed HPE and how we took a disciplined approach to capitalize on the multiple industry disruptions and turn them into value for our shareholders. Today, HPE is a vastly different company than it was 10 years ago. We are leaner, more innovative, with a differentiated portfolio, positioned to capitalize on the growing opportunities ahead. With the recent completion of our Juniper Networks acquisition, we move forward with renewed excitement. In this new chapter, I'm focused on accelerating value for our shareholders through higher profitable growth and increased capital returns. HPE was created with a bold vision to redefine and modernize enterprise IT for the era of digital transformation and the new age of insights. We architected a cohesive strategy aligned to the IT megatrends and implemented a series of strategic initiatives to reposition the company for the future.
We intentionally focus on what will differentiate us across networking, cloud, and AI through organic innovation and targeted strategic acquisitions. We simplified our company, improving our cost structure by re-engineering our processes and incorporating automation, and we never lost focus on our culture, including recruiting and developing the best talent to compete and win in the market. As a result of these deliberate moves, we significantly increased the relevance of HPE in the market, delivered solid shareholder value, and we set the foundation for what comes next. Now we turn the page to HPE's next chapter, one that unleashes HPE as a leader in high growth, high margin networking market. Our strength in networking will enable HPE to increase our participation in the AI and cloud markets with a stronger and stickier value proposition. HPE strategic priorities over the next three years are clear.
We will build a new networking industry leader, capture AI infrastructure growth profitably with a focus on sovereign and enterprise customers, accelerate our high margin software and services growth through our HPE GreenLake cloud, capitalize on unstructured data market growth with our own IP through Alletra MP storage, and drive customer transition to next generation server platforms. As we pursue these strategic opportunities, we will continue to focus on improving our operating leverage, and we are committed to delivering annual run-rate synergies by 2028 of at least $600 million through the Juniper integration and at least $350 million from our Catalyst initiatives, which are designed to accelerate revenue growth while driving structural cost savings that will allow us to return to 2x EBITDA leverage range by the end of fiscal year 2027.
We will generate more than $3.5 billion in free cash flow by fiscal year 2028 and will accelerate capital returns through higher dividends and increased share buybacks. I am confident in our ability to deliver on these strategic and financial priorities, which position us for success. HPE is playing to win in three strategic IT markets: networking, cloud, and AI. Each one serves as an essential building block for modern IT. These markets are all growing at an accelerated pace, driven by advancement in AI and the swift expansion in data center build outs. We anticipate the overall total addressable market across our portfolio will increase to over $1.1 trillion by fiscal year 2028. Our portfolio is well positioned to capitalize on all three growing markets. I believe the networking market will reach a significant inflection point, becoming the core foundation of the new AI technology stack.
This will require faster innovation to keep up with the disruptive innovation already experienced, we already seen in the server and storage markets through the acquisition of Juniper Networks and our unique expertise in data center scale. The Rascal solutions that we developed, HPE is well positioned to capitalize on the growth of the AI infrastructure in service provider, sovereign, and enterprise customer segments, and let me tell you how. AI at the core requires vast amounts of accelerated computing and data. Our AI systems are leading the market today with very strong solutions. Today, HPE ProLiant servers and Rascal solutions deliver sustainable accelerated computing for training and inferencing deployed at scale across the market. A unique fanless direct liquid cooling innovation is a point of differentiation for HPE.
Our global manufacturing services footprint provides us with a unique ability to compete around the world, and we complement this expertise with our market-leading supercomputing portfolio of HPE Cray EX solutions and systems, enabling the world to solve some of the biggest challenges through simulation, modeling, and now the accelerated adoption of AI. These capabilities are particularly important for sovereign and Neo cloud customers. Data fuels the development of AI models and applications. Unified high-performance storage accelerates access to insightful data, directly improving AI model outcomes, and as a result, organizations need scalable, integrated data solutions to expand AI initiatives. Our HPE Alletra MP Storage provides enterprise customers with a scalable, cloud-native, and AIOps-driven platform for structured and unstructured data, helping them reduce both CapEx and OpEx expenses. There is cloud. AI workloads are the true definition of hybrid. Hybrid and multi-cloud control are essential for flexible, cost-effective AI deployments.
Our HPE GreenLake Cloud enables enterprise customers to manage AI workloads across multi-cloud environments, simplifying operations and ensuring compliance while reducing operating cost through a unique agentic AI approach. Lastly, our financing and asset lifecycle capabilities set us apart. By providing enterprise customers with financial capacity for AI investments through sustainable approaches, HPE lowers the barriers to AI adoption, and again, our recent acquisition of Juniper Networks positions us to be the clear choice in networking, offering the industry-leading, secure, AI-native portfolio of networking spanning campus and branch, data center switches, and wide and wide area routing infrastructure. There is a clear advantage to such an expansive and integrated portfolio for our customers. We had a well-developed vision for the combined portfolios of Juniper Networks and HPE Aruba Networking when we announced the acquisition in January 2024.
Our market thesis has been further validated since then as customers ramp their AI infrastructure investments further and faster. Customers are attracted to HPE's open architecture and distinctive approach across the entire IT stack. Before the acquisition, we already had a strong portfolio, server, storage, networking, and cloud offerings along with global scale through our go-to-market reach. Now we are combining all those strengths as well as our extensive partner ecosystem with Juniper to create an IT powerhouse. I want to reaffirm our confidence in the success of this integration. My confidence is grounded not only in our rigorous governance and execution framework, but also in our proven success in integrating acquisitions that drove and continue to drive now strategic growth.
Whether it was from Aruba to Cray to Silver Peak to Morpheus to OpsRamp, we have consistently demonstrated our ability to bring companies together, unlock both revenue and cost synergies, and accelerate portfolio innovation. Let me be clear, Juniper will be no exception. We are driving this integration forward with a dedicated Program Management Office which is led by our Chief Operating and Legal Officer, John Schultz. John and the integration team offer support and oversight and track commitments and milestones around combining functions, operations, and go-to-market. An Executive Steering Committee oversees the Program Management Office and reports directly to the Board of Directors, who established an integration committee to ensure strategic oversight and accountability at the highest level of our company.
We have strong collaboration and cultural alignment across the teams with key talent from HPE and Juniper working as one team, already delivering on integration decisions and jointly engaging with customers. These early signals reinforce our belief that we will deliver at least $600 million in annual run-rate synergies by 2028 as we have committed to shareholders, which many are here in the room. Together, HPE and Juniper will disrupt the status quo in the networking market and unlock even greater value for our shareholders. Now to speak more about how we will deliver on our networking vision, we will welcome Rami Rahim who leads our new HPE networking business. Let's see what some of our customers have to say.
We live in incredible times where everything changed overnight. The use of AI is exploding. Hybrid cloud is a key enabler to deploy AI, and the core foundation of hybrid cloud is networking. By combining HPE and Juniper Networks, we have the potential to do some amazing, amazing things in the market. The world today needs a strong innovator at scale, and that is the power of this combination. This is once in a lifetime moment where you can really change the whole game of your infrastructure. Your solution strategy, the way you serve. Your business, the way you deliver value to your customers. The opportunity we have in front of us is enormous. To truly shake things up and innovate in ways that our customers and our partners have never seen, the time is now. It's helped us narrow the amount of. Physical devices we need and put us. Into a place where we can now grow on demand. We can see our entire infrastructure in one location, but more importantly, we can automate it. With HPE GreenLake and HPE Aruba Networking. We will set a new stage for enterprise networks globally. This is absolutely a game changer for us. It's transformative to the company and will be transformative to the world. The future belongs to the fast. Now we can turbocharge it by being together.
Thank you. Thank you very much and thank you, Antonio. Good afternoon everyone. It is exciting to be a part of HPE and to be here with all of you. Today I am privileged to lead the new HPE networking business and I'm here today to give you a clear view of our strategy for growing revenue, securing market leadership, and ultimately disrupting the networking industry. Now, as I share our plans, first I will talk to you about the market opportunities we are pursuing. In this new AI era, networking has become mission critical and as a result, we expect to capture market share in key network product segments that are both large and growing. Next, I will walk through our strategy for building a highly competitive, highly profitable networking business powered by the combination of HPE and Juniper Networks.
Third, I will highlight the unique and competitive advantages of our secure AI native networking portfolio and how our innovation and go-to-market scale can accelerate value to shareholders. Finally, I will discuss the results we intend to deliver and how we will track our progress over the next three years. Okay, let's dive right in. As Antonio mentioned, at HPE, we focus on three essential building blocks for modern IT: Networking, Cloud, and AI. In the era of AI, modern high-performing and integrated infrastructure is crucial, and the role of networking cannot be overstated. Connecting infrastructure, applications, users, and data more securely and efficiently is at the core of cloud and AI performance. Whether it's connecting people or scaling up and out to connect thousands of GPUs, networking is the foundation to delivering the business outcomes our customers need.
Legacy networks lack the necessary scale, reliability, intelligence, and automation needed for the modern AI era. Network performance is just no longer about speeds and feeds. It is about delivering a best-in-class experience to network operators and end users, whether those end users are people or AI agents. To achieve this kind of performance, organizations need a network that is purpose built with AI and for AI. When I say with AI, that means leveraging AIOps to simplify network deployment, stay ahead of security threats, and enable proactive troubleshooting to streamline operations, lower cost, and deliver an exceptional user experience. When I say for AI forming networks that minimize AI training and inferencing times, maximize GPU utilization and secure sensitive AI workloads and data. 30 years ago I was fortunate enough to have a front row seat as the Internet exploded to what it is today.
I am sincerely more excited today than I was about the opportunity back then because cloud and now AI have changed everything, creating new and even bigger growth opportunities for all of us. We see that growth opportunity in our overall networking TAM, which will grow to $169 billion by the year 2028. With HPE's acquisition of Juniper Networks, we now have the entire networking technology stack. We have the talent and the go to market scale to win in this growing market segment. Our vision is to deliver unparalleled secure AI native networks that empower our customers to meet their technical and their business objectives. To fuel this vision, my key priorities for the next three years are as follows. First, we will drive increased adoption of our highly differentiated AIOps solutions to capture, share and enhance profitability in AI for networks.
As our leading AIOps platforms come together and become even smarter, we will realize this vision of a self driving network which will increase our differentiation and drive both revenue growth as well as margin expansion. Second, we will tap into the large networks for AI investments that are happening in our industry to grow revenue and deliver shareholder value. Here, our greater R&D scale will enable us to innovate faster across networking silicon systems and software to not just participate in the AI data center market opportunity, but to be a leader in it. By providing a complete solution including routers, data center switches, firewalls, servers, storage, you name it, and the services all around it, we will win with our customers and we will grow our revenue. Third, we're going to claim market leadership in key areas where the network and security are converging.
Our approach to integrated security means that we deliver best in class efficacy, breadth and flexibility, giving customers supreme confidence in our solutions. As a result, we're going to be able to grow mindshare and market share. We plan to execute these strategic priorities in alignment with four distinct areas of the market opportunity where we have a competitive advantage that will accelerate value for our shareholders. Let me address each of them. I'm going to start with one I know is top of mind, camps and branches. This area represents a significant market opportunity for us as both the largest segment of the networking TAM and also the largest part of our business. Today we have strong momentum and recognition from the market with 18% market share and a robust ARR growth now exceeding $1 billion.
Our track record of innovation in the camps and branch is honestly second to none. We were first with a virtual assistant that uses generative AI conversation interface to solve problems using natural language queries. We were the first company with agentic digital twin technology that finds and fixes problems without users having to be present. We were among the first to introduce Wi-Fi 7 access points and have seen triple-digit growth this year in that market. We were the first to introduce the concept of a self-driving network, which is a game-changing approach to network performance measures focused on agentic AI and auto remediation. Now, as we build on this track record of innovation, there are three powerful reasons why our momentum is accelerating both with our customers and against the competition.
The first reason is that we have two industry-leading network AIOps platforms, Juniper Mist and HPE Aruba Networking Central, both with proven results delivering up to 90% reduction in trouble tickets and on-site service calls. Each of these platforms has unique and compelling strengths that actually complement each other. For example, Mist was purpose-built for public cloud deployments, while Central excels at private cloud and on-prem deployments as well as virtual private cloud architectures. Together they support every deployment option our customers want. Over time we will unify the experience of these two platforms by cross-pollinating features and capabilities of each onto the other. In so doing, we will protect our customers' investments while accelerating innovation and all at greater levels of engineering efficiency.
We are already moving really quickly to bring the best of both of these platforms together to all of our customers, and I look forward to showing exactly what that looks like at HPE Discover Barcelona in December. The second reason for our momentum is that we have one of the industry's largest data lakes. Our two platforms have been delivering powerful AIOps capabilities now for more than 10 years. We've been collecting anonymized network user experience data from billions of connected devices over that entire time frame. This gives us a very rich pool of data, which grows by more than a trillion telemetry points every single day. This means that our AIOps get smarter and more effective every single day, creating more powerful competitive moats around our solutions.
Third, we have consistently embraced forward-looking AI strategies, evolving from traditional prediction engines to gen AI capabilities like optimized search and chatbots, and now agentic AI, paving the way for a truly autonomous network. This progression reflects our ability to anticipate and lead industry shifts, delivering cutting-edge solutions today while positioning ourselves for future leadership. Our momentum is only accelerating, and the market is taking notice. In the 2025 Gartner Magic Quadrant for Wired and Wireless, both HPE and Juniper Networks were named leaders, while the long-standing incumbent was not. This milestone signals a significant shift in the industry and validates our growing market leadership. The next strategic market growth opportunity we see is in routing infrastructure solutions. Some people might know it as wireless LAN or WAN. AI adoption is driving unprecedented traffic demand not just within the data center but between them.
This presents a clear opportunity, as we have innovative solutions that drive compelling scale, performance, and power efficiency advantage that our customers need. This surge in AI traffic and AI training and inferencing is creating a massive data center interconnect, or DCI, opportunity with bandwidth demand expected to grow up to 6x in the next five years as enterprises and hyperscalers race to link GPU-dense facilities with low-latency, high-capacity networks. Furthermore, cloud providers are rapidly expanding their Direct Connect offerings, with some now delivering up to 400 Gbps of dedicated bandwidth per link. These cloud capabilities are enabling enterprises to move massive AI workloads between on-prem infrastructure and the cloud with predictable performance and low latency. Our HPE networking routing solutions, such as PTX for DCI and MX for Direct Connect, use our proprietary silicon offerings and our Junos operating system.
Our differentiated innovations were custom-built for scale and performance, uniquely positioning us to capture share as this space evolves. In addition to the AI opportunity, of course, we will continue to take advantage of technology upgrade and refresh cycles among service providers to drive revenue at healthy margins in that customer segment as well. Third is the data center market, and this, of course, represents one of HPE 's networking fastest growing opportunities today. Customers have varying data center needs on their location. Industry growth trajectory depends on many different factors, so one size fits all data center solutions are just not going to cut it. We will win in this market by pursuing both traditional enterprise data center opportunities and the fast growing AI data center market.
For traditional data centers, we will leverage our secure AI native networking platform for campus and branch and cross sell into the data center space. In addition, we will integrate into HPE 's hybrid and private cloud enterprise stacks to simplify deployment and day-to-day operations across compute, storage, and networking. Data center builders and operators need efficient management solutions like Abstra to meet their operational and regulatory requirements, their cost objectives, as well as their security imperatives. Abstra is truly unique in that it is the only solution that supports multi-vendor data center switching. It offers deep, actionable insight into networking events and its intent-based approach. It just abstracts away complexity from operators while ensuring that data center is always running as expected. For the AI data center market, success requires intimacy with cloud builders, including NEO Clouds and sovereign cloud providers.
Here, agility and speed of execution are absolutely key to success, and over the last couple of years, we have demonstrated an ability to do just that. We were first to market with an 800 Gb ethernet switch from our Q Effect series. We've landed meaningful front end and back end AI cluster wins from small to multi hundred thousand GPU cluster networks, and as we look to the era of 1.6 Tbps ethernet interfaces, we expect to continue our leadership. Liquid cooling is going to become critical for AI data center networks in the near future, and we will be able to incorporate our industry leading direct liquid cooling technology from HPE' s Slingshot products into our AI switching portfolio. HPE's experience and market leadership in the AI data center deployments around the world has already started opening doors for more networking opportunities.
Some cloud builders, especially sovereign clouds, prefer to work with a single technology provider for all aspects of their AI data center builds. HPE's ability to combine compute and networking into end-to-end data center solutions with expertise and services creates a big and growing opportunity for HPE networking. Network security now cuts across all of these markets and presents our fourth area of opportunity where we have a competitive advantage. The philosophy around network security is changing as customers demand security that's not bolt-on, it's built-in. Our objective in security is to strengthen our solutions, especially in the campus and branch and in the data center markets. As we execute our network and security convergence strategy, we will align ourselves to a growing trend of UZTNA, or Universal Zero Trust Network Access, that many enterprises are starting to embrace.
This advanced security framework applies zero trust principles consistently across all environments, whether users are remote, on-prem, hybrid, or using unmanaged devices like IoT. This is an emerging trend and we have a distinct advantage to lead in this space as we bring together our SASE, NAC, firewalls, and AIOps to create a complete end-to-end solution. By integrating security directly into the network, we protect our strategic solutions from standalone security players that are entering into the networking space and we will tap into a nearly $33 billion and fast-growing security market through an integrated approach and attach sales motions. We have a differentiated technology portfolio, we have market momentum, and we have the innovation prowess to bring this vision to life. Perhaps more importantly, we have the team we need to deliver on every promise we are making here today.
I have been incredibly impressed with how well our two teams have come together over the last few months. I have immense confidence in how we will continue to serve our customers and drive growth in our business as one networking team. Speaking of customers, HPE networking has some of the best customers in the world across practically every industry. Whether it's customers like Carnival Cruise Line, which deploys solutions from across the full HPE Aruba Networking and HPE Juniper Networking portfolio, institutions like University of Notre Dame, or convenience retailers like 7- Eleven. All our customers have one thing in common, and that is that their networks are absolutely mission critical to their organizations. I recently had the opportunity to see one of our customers' deployments in action at the 2025 Ryder Cup.
I also have to tell you that I'm a better technologist than I am a golfer, but I found that both aspects of that event were actually pretty impressive. With an integrated solution including HPE Networking, HPE Private Cloud AI, and a custom-built operational intelligence dashboard, we created one of the world's largest temporary smart cities on Long Island. Our solutions powered the tournament connections for everything from digital media to retail operations to course management to the fan experience for more than 250,000 spectators. Take a look.
The Ryder Cup without doubt is the biggest event in the golf industry. 250,000 spectators on the course watched by over 15 million people around the planet. It is real drama. Our ambitions on the Ryder Cup is. To be the most advanced, innovative technological platform that exists in the sport industry. The magnitude of the buildout of. This year's Ryder Cup is the largest that's ever been done. We get such a short period of. Time to maximize the revenue generation. The technology that we develop with Hewlett Packard Enterprise enabled us to be efficient, drive out costs from our business from hybrid cloud to private cloud AI through to HPE networking, using the data to drive the insight to drive the intelligence, principally to drive efficiency, particularly on the operational delivery. Across the Ryder Cup. Having the data at our fingertips is critically important. We have to make decisions very quickly. Having that live input is essential to operate efficiently. Enterprise bring to me such depth of portfolio, incredible expertise that allows me to deliver some of the most innovative solutions that exist in the sport industry today.
To help us deliver. Yes, okay, I love that video too. Look, to help us deliver on this kind of customer success, we have a robust and engaged partner community through which we do about 90% of our sales. When we closed this deal, there was only around a 10% overlap in partners between HPE, Aruba Networking, and Juniper Networks. We now have more than 47,000 partners around the world selling our HPE networking portfolio. Harmonizing partner programs and driving our combined product portfolio through a broader partner community represents a compelling source of revenue synergies for us. Additionally, as an accelerator to enable our partner community and capture new customer wins, we are harnessing the power of HPE Financial Services to make it easier for our customers to choose us.
HPEFS will help our networking customers and partners upgrade their technology at the pace and with the investment structure that they desire. Look, our networking segment is not just a growth engine, but the most significant contributor to HPE's financial strength. Over the next three years, we are committed to delivering strong results which will translate to enhanced returns for shareholders. This means that we will accelerate the adoption of our AI infrastructure native networking solutions as we pursue customers across every major industry. Accelerating customer adoption will enable us to achieve global market leadership in key segments including wireless LAN and data center networking, and we will deliver revenue growth with healthy margin expansion. Through this margin expansion, we expect to increase networking's contribution to HPE's non-GAAP operating profit to nearly 60% by fiscal year 2028.
HPE Networking was created to give the industry a new leader, a leader with a bold vision and a clear path to success. We have an immense opportunity to disrupt the status quo in networking and deliver unprecedented value for our customers, our team members, our partners, and our shareholders. I'm excited to be leading this business. I am confident we have the right strategy and the right team in place to capture the opportunity ahead. I want to thank you all for your time and I look forward to sharing updates on our success with all of you as we move forward. Thank you.
Thank you, Rami. We are excited about the momentum we are building together within Networking and across the rest of Hewlett Packard Enterprise. Thank you for joining us, Rami. Networking has become truly mission critical not just for connecting infrastructure, data, and users, but to power the next wave of AI-driven innovation across every industry. Networking alone is not enough. The future of IT will be built on the seamless integration of Networking, Cloud, and AI. Let's discuss Cloud and AI next. Our Cloud and AI strategy positions HPE to create value across several rapidly growing parts of the market, and our key priorities are: first, we will capture AI infrastructure growth profitably with a focus on sovereign and enterprise customers. In sovereign, we will leverage our supercomputing customer base and decades of expertise to capture the incremental AI business.
With enterprise customers, we are pursuing market share gains with our industry-leading turnkey AI factory solutions, including HPE Private Cloud AI and AI factories that scale at scale with our AI servers. In addition, we will accelerate high-margin GreenLake software and services growth, including virtualization, with expanded offerings to grow our profitable HPE Cloud Ops software suite. We are also scaling our go-to-market investment in Morpheus, Enterprise OpsRamp, and Zerto. Next, we will capitalize on the unstructured data market growth with our own IP through our unique Alletra MP storage platform. Finally, we will drive the enterprise transition to the next generation server platforms like HPE ProLiant Gen 12 while balancing units and profitability with higher services attached. Let's start with how HPE plans to profitably capture the tremendous growth opportunity that AI presents today.
Our AI offerings position us competitively across customer segments, including AI model builders and service providers. Sovereign and enterprise model builders and service providers comprise a significant portion of the market today, but deals in this customer segment are often very large, highly competitive, with limited services attached, and at very low margin. We continue to be very selective in pursuing these large-scale AI deals, ensuring each of them meet our working capital and profitability criteria. Our go-forward strategy in this market is to lead with Networking as we continue to look for opportunity for growth. However, the halo effect that comes from these high profile model builder deals helps us drive enterprise server and private cloud business, so it continues to be a very important part of our mix. The sovereign AI market is growing at an accelerated pace and diversifying.
For example, IDC research reveals that more than 50% of organizations in Europe are either already using sovereign clouds to build AI solutions or plan to do so within the next 12 months. HPE is uniquely positioned to capitalize on both the rapid expansion of sovereign and NEO cloud AI data centers driven by regulatory mandates and national security priorities. Our decades of HPC leadership, trusted government relationships, and our ability to deliver air gap private cloud solutions give us a unique competitive edge in enabling compliant, secure national AI infrastructure. Sovereign entities are responding to these advantages. An example of that is in Q3 fiscal 2025, sovereign AI orders grew 250% quarter- over- quarter, underscoring strong momentum in this emerging segment. In recent months, we inaugurated the fastest AI system in the U.K. It's called Isambard- AI with the University of Bristol and the U.K. government, and with Orange and the French Armed Forces Ministry, we inaugurated Asgard, Europe's largest classified AI supercomputer. In the U.S., the University of Utah plans to invest $50 million in AI, which includes NVIDIA AI computing by HPE. The build out is forecast to increase the University's computing capacity by 350%. We are leveraging our competitive advantage to accelerate this momentum and our leadership. HPE's clear supercomputing leadership and expertise give us a distinct advantage in AI. Six of the top 10 fastest supercomputers in the world have been built by HPE on HPE Cray EX Technology. In fact, along with the U.S. Department of Energy, we launched the world's top three fastest supercomputers, El Capitan at Lawrence Livermore National Laboratory, Frontier at the Oak Ridge National Laboratory, and Aurora at the Argonne National Laboratory.
In addition, the supercomputer segment enables us to pioneer new advanced technologies such as industry first 100% fanless direct liquid cooling architecture, which enables more efficient operations for computing at scale. One interesting factoid is that we already have direct liquid cooling systems deployed in more than 20 countries around the world, and the adoption continues to increase. HPE offers a compelling value proposition for enterprises accelerating their AI initiatives. As AI adoption scales, HPE is uniquely positioned to lead, combining advanced infrastructure, integrated services, and strategic partnerships. To meet this growing demand, many enterprises are turning to HPE in response to the urgent need for scalable, easy to deploy AI infrastructure. They do not want to take time or resources to custom build and integrate. Our HPE Private Cloud AI solution is redefining enterprise AI deployment.
This unique offer is different from the traditional architecture offered by our competitors. It is a fully pre-integrated solution that's ready to run in just a few steps. It is available both as a cloud-managed and air-gapped solution, making it ideal for enterprise customers with unique data compliance requirements, including sovereign use cases. A key strength of our Private Cloud AI solution is the comprehensive integration of high-margin HPE software and storage. The platform comes equipped with NVIDIA Accelerated Compute with our HPE ProLiant servers. The system also incorporates the HPE Cloud Ops software suite to simplify operations and NVIDIA NIM microservices to bring a true plug and play experience to users.
We launched this last year and it was already named by IDC as a leader for private AI infrastructure services systems. Our Private Cloud AI has now doubled new customer acquisition every single quarter to Q3 2025, building a pipeline that's bigger than $500 million in just the first year. As we look ahead, we anticipate substantial growth driven primarily by more enterprise adoption. A great example is Carbon3AI. A U.K.-based sovereign AI platform provider is adopting HPE's Private Cloud AI solution, which they plan to operationalize now in less than six weeks. This will enable immediate revenue recognition and provide cost control and data sovereignty, which were critical factors when Carbon3AI made their decision. For customers that require AI infrastructure that offers greater scale than our Private Cloud AI does today, we offer AI factories at scale. These tailored solutions combine high-performance compute with advisor and support services.
Customers include one of the top three retail banks in the U.S., several other financial services firms, and leading automotive manufacturers that rely on HPE AI servers to accelerate development of driver assist systems. We are leveraging our manufacturing services capability to win in global markets. For example, we have deployed one of the largest NVIDIA GB200 clusters in the world with an AI model builder pursuing an extensive large language model data center build out in Japan. Our customer KDDI will open a major data center to support startups and enterprises in developing AI applications and training large language Japanese models. Another example, Telus, one of Canada's most innovative communication technology companies, is leveraging our solutions and advanced computing infrastructure to launch Canada's first fully sovereign AI factory.
This operational facility is delivering cutting edge AI compute capabilities, empowering organizations to build breakthrough AI solutions while maintaining complete control over the data and innovations. Sovereign and enterprise AI customers segment now represent more than 50% of our cumulative AI orders book while enterprise AI orders have grown year- over- year in every quarter since the beginning of fiscal 2020. Now shifting from AI to cloud, over the last two years we have expanded our market leadership in the hybrid cloud segment by introducing HPE GreenLake cloud native innovations in highly attractive growing markets.
We are focused on building a high growth, high margin hybrid cloud business with an emphasis on leveraging our intellectual property and core differentiators at the center of our strategy. HPE GreenLake is the leading cloud delivering a unified platform experience that allows enterprises to simplify IT, reduce cost, and transform faster as they incorporate their AI deployments. Today we have 44,000 HPE GreenLake customers, an increase of 7,000 new logos, up 20% from a year ago. We help these customers manage more than 5.8 million devices and over 8 exabytes of data, generating over $2 billion in ARR. In June we introduced HPE GreenLake Intelligence, our industry first agentic AI ops framework which is multi cloud and multi vendor. HPE GreenLake Intelligence uses agentic AI to simplify the IT operator's experience across storage, networking, compute, runtime, and applications.
The power of AI for networks that Rami discussed earlier will accelerate the vision of HPE GreenLake, which is already powering tens of thousands of HPE Aruba Networking users with cloud native and AI driven networking and observability services. One of HPE GreenLake's competitive advantages is its software. In June 2025, we introduced HPE Cloud Ops Suite, which is a software suite that includes cloud management orchestration from Morpheus, observability from OpsRamp, continuous data protection from Zerto, and now runtime from Morpheus VM Essentials. It is complemented by our HPE GreenLake Intelligent Agentic framework, which provides an integrated AI-driven experience. We are the only vendor that provides these capabilities across multi-cloud and multi-vendor environments. A great example is the state of Vermont. This proved to be a valuable differentiator unmatched by our competitors.
OpsRamp provides the observability and automation backbone for managing Vermont's hybrid cloud infrastructure, streamlining operations and reducing complexity. OpsRamp also supports intelligent monitoring and incident response, allowing Vermont's IT staff to shift from routine tasks to higher value projects. In addition, OpsRamp natively integrates with leading ITSM platforms such as ServiceNow for effective IT trouble ticketing management. Looking ahead, we will accelerate higher margin GreenLake software and services growth with expanded offerings and scale out our go-to-market investment in Morpheus Enterprise, OpsRamp, and Zerto. In addition to deploying AI, enterprise customers around the world are also reevaluating their virtualization needs driven by the recent Broadcom bundling and price changes. In fact, 74% of Gartner Peer Insights community members polled are actively looking to VMware alternatives. In response to these market conditions, we enhanced our private cloud offerings by integrating Morpheus software as a standard component.
This addition not only enables seamless multi-cloud orchestration between public and private clouds, but also supports a broader partner ecosystem, empowering customers with greater choice and flexibility. HPE Morpheus VM Essentials were designed to help customers adopt as an alternative virtualization solution without compromising on the enterprise-grade reliability. Our solutions deliver significant value by reducing customer licensing cost by up to 90% compared to VMware. Danfoss, a Danish manufacturing company, is our first customer to implement this offering. With the company's existing hypervisor contract about to expire, HPE introduced HPE VM Essentials as an alternative, helping Danfoss avoid costly renewals and also gaining at the same time a self-service cloud experience. Danfoss can shift now from what we were doing today to more an as-a-service model because of our standardized private cloud edge solution using HPE.
HPE VM Essentials is now being rolled globally at their 65 remote sites, and we plan to continue to increase customer value by integrating now intellectual property from Juniper Networks switches and the software-defined networking that Juniper Networks brings to HPE with all our HPE Private Cloud solutions. This tight integration exemplifies the power of uniting cloud, AI, and networking capabilities through our HPE GreenLake cloud, reinforcing our better together value proposition for our customers. The growing adoption of AI applications has made data management a key focus for enterprises. The effectiveness of AI is directly correlated to how well customers can access and manage distributed data across their operations, and as a result, there is a growing demand for HPE's data fabric, HPE storage, and data protection solutions. Our HPE Alletra MP is the world's fastest growing all-flash structured data platform.
It pioneered the world's first disaggregated storage architecture and offers a true cloud-native experience through HPE GreenLake. These unique features resonate with customers and have resulted in HPE taking market share for four consecutive quarters. In fact, we gained three points of market share in block storage year over year in the second quarter of calendar year 2025, growing seven times faster than the market in all-flash block. The HPE Alletra MP platform now also supports unstructured data and cyber resiliency use cases, broadening the presence in the rapidly expanding AI market while complementing our networking and server offerings, and our customers are responding. A great example is Qualitas, Mexico's largest auto insurer, which knows that every second of downtime threatens millions of customers and interactions and claims.
They work with us to scale compute and storage capacity to support the data analytics and ensure high availability for their mission-critical workloads, helping reduce operating costs without sacrificing performance. It also enabled them to build a flexible foundation for AI and hybrid cloud adoption. The HPE Alletra MP architecture delivers unmatched performance, service levels, agreements, and total cost of ownership for Qualitas. With the HPE Alletra MP success demonstrated to date, we are now ready to take the next step in the transition of our storage portfolio. Historically, our storage portfolio sales rely on a combination of both own IP offerings and third-party products. Now we are evolving our model to focus on HPE-developed storage solutions, reducing dependence on third-party products except where we collaborate on very specific differentiated customer solutions with a very few curated strategic partnerships.
This shift demonstrates the maturity and the strength of our storage offerings as well as the completeness of our portfolio, and we are confident with the investments we will continue to make in our own intellectual property, we are better positioned to support our customers through our unified architecture and customer experience. In our core server segment, we are managing profitability in units in a balanced way with a focus on volume growth and services attached to sustainability, profitability, and cash flow generation long term as we have demonstrated this year with the transition to our HPE ProLiant Gen 11. Each generation transition drives revenue growth through richer configuration and higher average unit prices. While we already begun the ramp into Gen 12, the rest of the transition will provide a few further tailwinds for our customers.
The transition to our newest servers has delivered significantly better computing performance along with real estate and power savings through rack consolidation and improved energy efficiency. In fact, when you upgrade any Gen 10 server from any vendor to an HPE ProLiant Gen 12 server, that upgrade yields 65% annual power savings and a 7 to 1 reduction in the data center footprint. Through our HPE Financial Services, we are enabling customers to unlock capacity through financing and IT lifecycle services to support capital required to modernize the server infrastructure. Finally, we will continue to expand our global manufacturing footprint, including new server production capabilities already in place in India and Saudi Arabia. This local manufacturing strategy gives us a competitive advantage to serve high-growth emerging markets, including the Middle East and Africa. HPE NonStop, we have a lot of products.
HPE NonStop is where customers turn to run the most mission-critical transaction workloads. We recently introduced our next generation of HPE NonStop servers, which is triggering now a refresh cycle for this very lucrative business. NonStop is vital for payment processing and fraud detection. NonStop is also categorized as an AL4 with uptimes of 99.999% or even higher. In the Financial Services vertical segment, six of the top 10 full service global retail banks use NonStop for high volume payment processing, ATM functionality, and core banking capabilities. Our next generation of HPE NonStop solutions we just introduced double the amount of memory and bandwidth, providing our mission critical customers with the latest technologies to maintain that uptime and improve business resiliency. Finally, quantum computing is an emerging opportunity that's rapidly gaining momentum.
Over the last few quarters, interest in quantum computing has started to translate into real customer demand specific for systems that tightly integrate quantum capabilities with high performance networking and computing. HPE Labs Networking, now with the amazing team that we built together with Rami and Supercomputing organization, have been at the forefront of innovation critical to this space, and their work has created opportunity for us. By leveraging this expertise and by taking a vendor agnostic approach to integrating third party quantum technologies with our networking and HPC software intellectual property, we will provide customers with options and they need to unlock the full potential of quantum. HPE has generated interest in these capabilities with several customers, and engagement is ramping quickly.
With the acquisition of Juniper Networks, we now have the ability to accelerate our role in delivering quantum scale out systems over time, and over the next few quarters you will see us demonstrating some of these capabilities and the progress we are making together with our customers and quantum partners. One key aspect of our portfolio that has only become more relevant in the AI era is HPE Financial Services, which accelerates adoption of AI through financing and IT lifecycle services as the captive finance arm for HPE and preferred financing vendor for HP Inc. HPE Financial Services help customers and partners transform their enterprises at the pace and investment structure they desire. We will leverage the unique value proposition of our technology renewal centers to expand our residual-based leasing portfolio to accelerate our customer adoption of the full stack of networking, cloud, and AI technologies.
HPE Financial Services' unique combination of financial discipline, IT lifecycle services, and experienced team that we have enables the business to deliver a consistent and steady return on equity of approximately 16%, operating profit of 10%, and predictable financing volume. While Financial Services is helping customers transform with flexible financing and lifecycle solutions, we will also take bold steps internally to transform how Hewlett Packard Enterprise operates, and this is where Catalyst comes in. Catalyst is a set of company-wide initiatives to make HPE faster, smarter, and more efficient. It goes beyond cost savings. Catalyst invests in AI and data platforms, better data governance, and smarter ways to manage information. The goal is to embed AI in everyday work, automate key processes, and unlock insights from data so HPE can focus on high-growth areas while remaining agile and competitive, with Catalyst helping us become more nimble and efficient.
Our differentiated portfolio and our talented team give me great confidence in HPE's next chapter. Before I turn it over to Marie, I want to sum up what we have discussed today so far. Our priorities for the next three years are clear. We will strengthen our market position in establishing leadership in networking, expanding profitable opportunities for growth in AI, boosting growth in hybrid cloud software and services through HPE GreenLake, capitalizing on the unstructured data market growth with Alletra MP, and guiding our customers towards advanced server technologies. We are committed to driving operating leverage by delivering approximately $1 billion in annualized structural cost savings to shareholders by fiscal year 2028. We will do this by realizing synergies from Juniper Networks integration and execution of Catalyst initiatives, both designed to accelerate revenue growth and generate substantial structural cost savings.
All these actions will enable us to generate more than $3.5 billion in free cash flow by 2028 and deliver greater shareholder returns through growing our dividends and share buybacks. As I look ahead to HPE's next chapter, our commitment to innovation and shareholder value has never been stronger. To share how these strategies will shape our future and deliver lasting impact, I'm delighted to welcome Marie Myers to stage. Marie will discuss how we are building a modern company and provide more details which all of you are waiting. I know that on how Hewlett Packard Enterprise will drive greater value for our shareholders through our financials. Marie, floor is yours.
Thanks, boss. Thanks for joining us today. I'm thrilled to speak with all of you during such a pivotal moment in HPE's journey as we embark on the next chapter of our growth. As many of you know, I joined HPE. It's hard to believe it was only January last year and I joined the company, as many of you also know, from HPQ, because I believed in the vision of this company. I saw substantial opportunity for HPE then and now I see even more. We have many new chapters of growth ahead of us and I'm excited to speak with you today about how we intend to set ourselves up to capture the significant opportunity available to us over the next few years. Now that we have closed the Juniper Networks acquisition, we are uniquely positioned to evolve HPE into a modern AI-led networking company.
We are transforming not only the profit profile of the company, but we are also fundamentally rethinking how we operate and how we work. By embedding AI across our operations, we are driving smarter, faster decision making and unlocking new levels of agility. This transformation is about building a more responsive, scalable organization, one that can adapt quickly and compete in what is an incredibly dynamic market. Moving forward, we will further simplify our operating model, helping us sharpen our focus and deliver more profitable revenue growth. As you heard from Antonio, accelerate greater capital returns to our shareholders. However, we will continue to invest in our portfolio while staying disciplined, as you would imagine, with our cost structure. To drive that profitable growth, we will create the operating leverage that supports meaningful earnings expansion over the next three years.
Today, I'm going to share our plan to achieve that acceleration of value. First, I'll walk you through a quick recap of our Q3 year-to-date 2025 results. Then I will share our progress on the transformation that we've been undertaking. Third, I will explain our go-forward shareholder financial framework that prioritizes generating more than $3.5 billion in free cash flow by 2028. This will enable HPE to reduce our financial leverage and increase capital returns. Finally, I know you. Let me begin with a quick recap of what we have accomplished so far in fiscal 2025. We have momentum in AI and you have seen a meaningful recovery in networking, which puts us on track for strong revenue growth through Q3 alone.
Top line revenue is up 14% year- over- year and we've worked hard to increase operating profit with a disciplined focus on margin recovery in the traditional server business, and our results are further supported by four consecutive quarters of operating margin expansion in hybrid cloud and a larger contribution from networking. We have also maintained strong non-GAAP earnings per share through disciplined cost management and operational efficiency, and today we are reaffirming our FY 2025 non-GAAP outlook. However, as previously noted, revenue may be impacted by lumpiness and deal timing in AI, and in addition, we are updating our Q4 GAAP diluted EPS guidance from what was a range of $0.50- $0.54 to $0.11- $0.15, and for the full year from $0.42- $0.46 to $0.03-- $0.07.
This reduction in our expectations is driven by certain unanticipated events that occurred after we issued our Q4 guide, including losses on disposition of certain non-business assets, acquisition-related accounting adjustments, updated forecasted tax expense, and other items. Free cash flow guidance remains unchanged at approximately $700 million, impacted by working capital dynamics and Juniper-related costs. We anticipate strong free cash flow growth over the next three years, and I will talk shortly about a 10% increase we are announcing in our dividend for common shareholders starting in Q1 2026. You have heard us talk about our Catalyst program and Juniper cost synergies throughout this year. I'd like to take some time now to describe why this is so critical to simplifying and transforming HPE while improving our operating leverage.
Our teams are hyper-focused on achieving structural cost savings through Juniper-related synergies and Catalyst initiatives, and our two targeted sets of actions are designed to increase productivity, capture efficiencies, and unlock operating leverage that will drive long-term sustained profitability. Together, they're expected to deliver approximately $1 billion in structural annualized run-rate cost savings by 2028. This includes at least $600 million from Juniper and another $300 million from Catalyst. We expect these initiatives will reshape us entirely, making HPE a much more nimble, cost-efficient company. Savings will benefit both the cost of sales and operating expenses while supporting investments vital for long-term sustainable growth. We established a strong Juniper integration framework and Program Management Office. Our integration plans are on track as we begin building the best-in-class networking business with Rami, and we remain confident in achieving both near-term financial synergies and long-term targets.
We began our integration process with corporate function overlaps, streamlining combined functions to eliminate duplicative roles and improve efficiency. While workforce reductions are always challenging, they are necessary to align resources and to drive productivity. Most integration-related reductions will occur in year one post-close, with some extending into year two and year three. For more complex integration areas, we will complete integrating our sales teams across regions by Q1 2026, and we will continue to align channel programs and sales tools through next year so that we can enable partners to sell the full portfolio and unlock further growth. In terms of product rationalization, we will take a very measured approach to preserve our customer experience. However, we will begin aligning R&D spend to the strategic priorities that Rami referenced earlier while converging product roadmaps and eliminating duplicate R&D projects.
We are leveraging the combined scale of HPE and Juniper Networks to unlock efficiencies in logistics, materials, pricing, warehousing, and repair services. We expect to begin realizing these synergies in the first half of 2026, and we are also deploying AI to further optimize material costs and refine our manufacturing footprint, including aligning headcount to our go-forward operational needs. Additional synergy savings will come from areas like real estate consolidation, IT optimization including software support services, and actually reducing reliance on third-party contractors. These efforts combined will deliver meaningful structural savings while preserving critical operational capabilities. Overall, we expect to realize about $200 million of Juniper-related cost synergies in year one post-close, with the balance of $400 million split evenly between years two and three. These synergies will require approximately $800 million of investment to achieve between 2026 and 2028.
Most of that is going to be tied to headcount, supply chain optimization, and portfolio rationalization. The bulk of the integration expense will occur next year. I have discussed Catalyst, our multi-year cost reduction and efficiency actions, with you since we launched it earlier this year. Catalyst is about fundamentally reshaping how we operate, reducing processes, creating capacity to reinvest in that growth that we've talked about, and most importantly transforming the company for long term competitiveness. I'm really encouraged by the progress we've made in just this year alone. We are just at the beginning of this journey to drive improved operating leverage for our company and frankly better experiences for our customers and all our team members. Catalyst is anchored on four strategic pillars: Workforce Transformation, Portfolio Optimization, Operational Excellence, and finally leveraging AI. First, Workforce Transformation.
It's by far our greatest opportunity and frankly we are redesigning our work to deliver better results with fewer resources. I shared earlier this year that our efforts alone had reduced the headcount at HPE to the lowest historical levels ever pre-Juniper, and frankly, we continue to proceed with simplification principles in mind, streamlining organizational spans and layers, centralizing roles, and frankly encouraging our teams to concentrate on innovation and high value activities. Second, Portfolio Optimization. We're sharpening our focus on strategic assets. For example, we exited a non-core business at the start of this fiscal year, divesting our communications technology group, and we will continue to evaluate our non-core assets in future, including pursuing smaller size divestitures to prune our portfolio. We're also actively reducing SKUs, continuing to trim our product offerings to improve speed and efficiency.
Third, with respect to operational efficiency, we are revisiting business operations and instilling greater discipline across our businesses. A prime example of that is what we've done with pricing. We've actually embedded AI and advanced analytics to improve deal visibility, reducing unnecessary discounts and maximizing margin capture across the entire portfolio. I'd like to highlight one of the recent projects. We're actually partnering with Deloitte. Big shout out to the Deloitte team that are here today to co-source several processes in my own organization that will greatly streamline operations and allow our finance team to move forward in a much more agile, AI scalable way. We are also driving flexibility through vendor optimization and real estate consolidation. Finally, the category I'm most excited about, AI. That's because it lets our technology-minded workforce operate smarter, faster, and with precision.
We are using AI as a transformation lever to free up resources across the company, making us much more agile in the market. We are building a robust internal data foundation and adopting an AI-first approach across the company. Let me share a few examples of where we're taking a fresh look at the work we do. Marketing content creators at HPE today are using generative AI to accelerate how quickly they develop smart campaigns and reduce production costs. In finance and sales, believe it or not, it's improving forecasting, reconciliations, and opportunity management to improve our financial insights. We have also been partnering with Deloitte to develop an AI platform to transform our financial operations. It's something we are commercializing through our own HPE Private Cloud AI infrastructure and making it available to other CFOs and customers.
As you can tell, we are enthusiastic about streamlining how we operate, centralizing functions, and using AI tools to drive speed, agility, and competitiveness. Finally, Catalyst will allow us to target at least $350 million in annualized run-rate cost savings by the end of 2027. We are executing well against our plan and expect to achieve our target of 20% of total savings by the end of 2025. As you know, that has a direct impact on both OpEx and cost of sales. As we disclosed earlier this year, these savings will require approximately $350 million in charges to fund workforce reductions. Catalyst is not only reshaping our cost structure and improving financial flexibility, but it's reshaping the way we work and it will provide us with the capacity to reinvest in growth.
I'm pleased to say we're on track to create a leaner, much more competitive company that will continue to deliver proper growth to our shareholders. Today we are also announcing a framework for simplified segment reporting and disclosures starting in Q1 of next year. Historical segments will, of course, be restated back to Q1 FY 2020. We will also provide visibility into each new segment. Going forward, I believe you will gain a clear line of sight into the growth trajectory of our portfolio while having the ability to compare much more easily across our peer set. This new structure aligns with the main markets in which we operate and where we see the greatest growth potential. It's simple, it consists of three segments. The first segment, Networking, includes Juniper Networks and Intelligent Edge. The second segment includes Cloud and AI, and it consolidates server, storage, and financial services.
Finally, our third segment, Corporate Investments and Other, will be largely unchanged. Networking will have a larger footprint within our results and, as you've heard, it now represents on a pro forma basis nearly 30% of our total revenue and over 50% of our total operating profit. Given our increased scale and strategic objective to establish ourselves as the industry leader in networking, we want to provide you with more detailed financials, and in this segment we will report quarterly revenue by customer groups, enterprise and service providers, as well as by product line, campus and branch, data center, networking, routing, and security. This will provide important visibility into our traction with key customers. It also allows you to see product line highlights for what we consider our core strategic and market priorities.
Annually, we will disclose our outlook for cumulative networking, for AI orders, and then for cloud and AI. This segment signifies the importance and the rise of AI as customers implement and scale these technologies. Incorporating HPE Financial Services within this segment demonstrates how financing is playing a key strategic role in helping enterprise customers adopt and consume these technologies. This underscores that our focus is simply not just about hardware, it's also about recurring revenue. Here we will report disaggregated revenue results as server, storage, financial services, and other, and the other sub-segment will also include non-IP storage. As you heard from Antonio, this is an area we are de-emphasizing as we transform our storage portfolio to focus on the success of our own HPE Alletra MP products and I will discuss more on that later. We will also share additional quarterly disclosures for server offerings.
We will disclose AI orders and revenue as well as AI customer mix, and within the storage category we will provide HPE Alletra MP orders, growth, and other financial services. We will provide return on equity to show the profitability we generate on investments that we make for our enterprise customers. Finally, we will continue to provide visibility into ARR, which is a key forward-looking metric for our long-term recurring revenue. ARR growth will moderate due to a combination of factors going forward. Given that we have built up such a large base of ARR, we will make it an annual disclosure on a consolidated basis. This new structure streamlines our reporting framework and aligns with how we operate our business on a day-to-day basis. I believe it also gives you all a much clearer picture of where we are driving growth and increasing value for our shareholders.
Now I would like to walk you through our strategic framework for shareholder value creation. At the core of this strategy is our free cash flow growth. It is the engine that powers everything we do. It enables us to reinvest in innovation, scale our business, strengthen our balance sheet, and return capital to shareholders as we enter our next chapter as a company. Our portfolio has been designed to win in networking, cloud, and AI, and the sustained profitable growth we envisage is supported by cost discipline, structural improvements, and capital efficiency. We expect this operating model will deliver higher, durable free cash flow over time. We are driving this free cash flow through three key levers. The first lever is profitable revenue growth. We are scaling in high-growth areas like AI services and networking, where we will transform our portfolio to deliver differentiated value.
There is also operating margin expansion through disciplined execution and cost reductions, especially those around Juniper Networks and Catalyst initiatives. We will deliver sustained margin improvement. Finally, we have balance sheet efficiency. We are managing our balance sheet with rigor and unlocking liquidity through tighter operational decision discipline. As Antonio discussed, our strategy to generate free cash flow of $3.5 billion is core to us and we expect to achieve that by 2028. We expect annual free cash flow generation to improve in 2026 and then continue to ramp over the next three years. We expect to generate approximately $700 million in 2025, rising to $1.52 billion- $2 billion in 2026 and increasing further to $3.5 billion or more than $3.5 billion in 2028. In FY 2026 we will incur the bulk of the Juniper integration expenses, Catalyst initiatives related cash costs, and higher interest and acquisition related debt.
Together these levers are building a stronger, more resilient free cash flow profile for us to execute on three financial priorities. First, we are accelerating repayment of Juniper related debt to ensure a strong financial foundation with our commitment to maintain an investment grade credit rating. Our goal is to reduce net leverage to 2x by the end of 2027 through both net debt reduction and EBITDA growth. We've already made progress with a $1 billion repayment post Q3. We expect our liquidity position to strengthen further as we grow free cash flow, pay down our obligations, and diligently manage our balance sheet. Starting in fiscal 2026 we are increasing our annual dividend by 10% to $0.57 per common share, reflecting our confidence in the outlook and commitment to deliver reliable income to common shareholders.
Lastly, we are managing dilution in the near term through 2027 and we expect to accelerate repurchases in the latter half of 2028. Today we are announcing a new $3 billion share repurchase authorization, bringing our total to $3.7 billion. As we generate increasing levels of free cash flow over time, we intend to return a significant amount to our shareholders. With our focus on reducing debt through 2027, we will return nearly all of our free cash flow to debt and equity holders over the next two years. Thereafter, we will target returning at least 75% of our free cash flow to our shareholders in the form of both dividends and share repurchases. Our focus on these core financial priorities and our new capital returns commitment will ensure we remain shareholder focused and committed to long term value creation.
Now let me walk you through our three year financial outlook. We expect consolidated revenue to grow at 5%- 7% CAGR on a pro forma basis including Juniper. In networking, we project revenue to grow at a 5%- 7% CAGR over the next three years on a pro forma basis in line with what Rami shared in terms of our estimate of the addressable networking market. Our enhanced scale and competitive positioning gives us confidence that over time we will outpace market growth in our key focus areas. While Juniper's current margins are below historical HPE networking business, we expect to offset dilution through cost synergies and scale efficiencies. As a result, we expect networking margins to reach 25%- 28% by FY 2028, driven by Juniper related cost synergies and operating leverage as we scale revenue and absorb fixed costs over a broader base.
In cloud and AI, we expect revenue to grow at a 4%- 8% CAGR over the next three years with margins in the 8%- 10% range. Let me provide you some additional insight on the strategic drivers for the segment. First, AI infrastructure, as Antonio mentioned, represents a significant opportunity as adoption expands beyond service providers and model builders into more profitable enterprise and sovereign segments. We are differentiating ourselves by maintaining our discipline in the pursuit of profitable growth, evaluating each opportunity to ensure the right balance between revenue and margin and working capital. As a result, AI infrastructure revenue growth may trail the broader market as we focus on free cash flow generation. Next, HPE GreenLake is central to our strategy and is driving over 75% of our $3.1 billion ARR in Q3 2025 coming from software and services.
Looking ahead, we expect ARR to grow approximately to $3.8 billion by the end of 2026. In storage, we are doubling down on our Alletra MP product strategy, which is significantly more margin rich. We are seeing strong traction for our Alletra MP platform, and this success enables us to pivot away from selling our non-owned IP, which tends to dilute margins. As a result, we expect margin improvement over time, positioning us for stronger long-term profitability. In traditional server, we are balancing units and margin while expanding our install base and services attach opportunities to drive further profitability. We expect traditional server growth will be fueled by ongoing IT modernization and data center expansion. Finally, our financial services business remains a key competitive advantage. Our disciplined underwriting approach allows us to finance high quality customers which deliver a strong return on equity of nearly 18% as of Q3.
Given these segment dynamics, we expect the overall company non-GAAP operating profit CAGR to be 11%- 17% on a pro forma basis, driven by higher mix of networking, a favorable mix shift within cloud and AI, structural cost reductions, and enhanced operating leverage. By FY 2028, we expect non-GAAP diluted EPS to grow to at least $3 and free cash flow to exceed $3.5 billion. Turning to our 2026 outlook, we anticipate revenue growth of 5%- 10% on a pro forma basis. Our revenue growth reflects the lumpy nature of AI transactions. Segment-wise, networking revenue is expected to grow at a low to mid single- digit rate on a pro forma basis, driven by continued momentum in campus and branch, SASE and security, and data center.
We also expect networking for AI demand will improve during the year, with cumulative orders increasing to $1.5 billion by the end of 2026, up from $600 million of orders generated during this fiscal year. Cloud and AI revenue is projected to grow at a mid single to low double- digit rate, reflecting solid execution in our server business as we capitalize on IT modernization and AI investments, maintaining our focus on profitable server markets and offsetting revenue declines from exiting our non-IP storage business. It's important to note that historical seasonal trends will not be indicative of future performance as synergies ramp throughout the year. We expect margin performance and net earnings to strengthen significantly in second half of 2026, making the second half significantly stronger than the first. From a profitability standpoint, we expect non-GAAP operating margin to improve year- over- year to a low double- digit rate.
This includes networking operating margin in the low 20% range, supported by top line growth and deal synergies. For cloud and AI, we expect stable margin of 7%- 9% for the year, driven by normalized margins in traditional servers, a short shift to owned IP and storage, and continued investment to scale our hybrid cloud business. Margin variability from AI deal mix and near term scaling costs will be managed prudently and from a cost perspective. Total costs will be up year over year due to a full year of Juniper contributions, but we expect and continue to expect that our cost initiatives will deliver the projected savings accounted for in our non-GAAP diluted net EPS guide. We expect FY26 non-GAAP diluted net EPS to be in the range of $2.20- $2.40, up approximately 20% at the midpoint of our outlook.
This reflects several items: growth in the core business, eight months of incremental contributions from Juniper, Juniper-related synergies, and Catalyst costs, offset partially by non-operational headwinds including approximately $650 million in OIE, and lastly a full year diluted net share count of 1.44 billion. Additionally, we are revising our non-GAAP tax rate to 14%, down from 15% due to the benefits from the Juniper acquisition. GAAP diluted net EPS is expected to be in the range of $0.57- $0.77, and we expect FY 2026 free cash flow of $1.5 billion- $2 billion, which includes $600 million worth of costs related to the Juniper and Catalyst programs. In closing, we have built a portfolio designed to win in the next AI era. Positioned at the intersection of networking, cloud, and AI, I believe HPE stands out as an exciting and compelling investment.
We believe our operational discipline and execution will drive profitable growth, and our cost initiatives will unlock meaningful operating leverage. Our business will enable us to realize robust free cash flow expansion expected to exceed $3.5 billion by 2028. This financial strength gives us the flexibility to invest in innovation, scale strategically, and deliver accelerated capital returns through consistent dividend growth and share repurchases. With a clear commitment to shareholder value, we believe that HPE is well positioned for long-term success. I appreciate this company's enthusiasm for the transformative steps we are taking to build our next chapter of growth. I'm privileged to work alongside Antonio and the rest of our leadership team as we enter the next phase of this journey. Thank you all. Now we're going to take a short 15-minute break. Thank you.
It's in our DNA, and when harnessed, it can change everything.
Ambition turns oceans of data into waves of possibilities. It transforms artificial intelligence into real-world impact, and one day the final frontier into our backyard. Unlock the promise of AI and whatever comes next. We're at a new threshold where human expertise is harnessing the unprecedented power of AI, opening the door to infinite. BMW reimagining sheer driving pleasure with a groundbreaking approach to electric vehicle design that transforms automotive manufacturing and is powered by data to advance performance and safety. BMW's global engineering team tests the EVs in diverse driving scenarios, from long distance to extreme weather. Using HPE GreenLake, data is processed from each vehicle on an edge computing platform, assisting the team in leveraging analytics and machine learning, engaging AI to deliver vital insights. All allowing for seamless collaboration, development, and a greater speed to market. This is what we unlock when intelligence has no limits.
You drive with better performance, more confidence, more sustainably. We drive a data foundation, advanced connectivity, faster innovation. You drive with joy. We drive towards an industry transformed. Unlock adventure. Collaboration. Joy. HPE unlocks ambition. three, two, one. The moment has arrived. And so have you. You're more than a guest. You're the main character. Surrounded by ocean.
Whatever you want is at your fingertips. Carnival Cruise Line is partnering with Hewlett Packard Enterprise to ambitiously redefine the vacation experience. Experience a floating city miles out at sea brought to life by HPE Aruba networking with seamless shipwide connectivity. From your first step on board, everything feels personal, responsive, and effortless. The crew is always on hand, quietly working behind the scenes to make you feel like a VIP.
Facial recognition enhances that experience.
Letting you board safely and smoothly to helping you capture your memories during the cruise. Making it all possible are two HPE powered data centers, essential technology when you can't afford downtime this far from shore. Shipwide connectivity gives the crew instant insights on everything from weather patterns to performance and security data that anticipates needs before they even arise. The crew can focus on those small, thoughtful details that make every guest feel truly cared for. Together, HPE and Carnival Cruise Line are expanding your imagination and how far travel can take you. Unlock fun. HPE unlock ambition. Today, countless times, all over the world, a plane will leave the runway carrying passengers across the horizon. The doors will open to new adventures, friendships, even new lives. Flight has freed us for endless possibilities.
Now aviation itself can aim higher to bring the world closer with more efficiency and less impact on the environment. GE Aerospace's ambition is to decarbonize commercial flight with jet engines that enhance performance, reduce fuel consumption and greenhouse gas emissions, and take the industry in a new direction. At Oak Ridge National Laboratory, GE Aerospace engineers use an HPE AI supercomputer to run simulations at close to a quintillion calculations per second to optimize components, testing countless versions to achieve their breakthrough design. This is what we unlock when intelligence science has no limits. While you explore navigating new landscapes, embracing new cultures, discovering new perspectives, we explore aerodynamics, acoustic physics, engine performance. You explore endless horizons. We explore the future of flight. Unlock exploration, determination, curiosity. HPE unlock ambition.
Football, soccer. The world's game. It's more than just a sport. It's a global phenomenon that unites billions. Every rule, every chant, every. They are more than just moments behind. Every moment a story. Behind all the data, a thing. Now, FC Barcelona is teaming up with Hewlett Packard Enterprise for a world class cloud platform. A networking solution that sets a new standard. FC Barcelona and Hewlett Packard Enterprise together, we're crafting an immersive experience where fans aren't just. Spectators, not part of the game. Together, we're building more than just a stadium. We're building the future of football. HPE. Powered by innovation, fueled by passion. Unlock pride. Unlock community. Unlock destiny. HPE. Unlock ambition. Ambition. It's in our DNA.
When harnessed, it can change everything. Ambition turns oceans of data into waves of possibilities. It transforms artificial intelligence into real world impact. One day, the final frontier into our backyard. Unlock the promise of AI and whatever comes next. We're at a new threshold where human expertise is harnessing the unprecedented power of AI, opening the door to infinite possibilities for changing our lives. Enter BMW, reimagining sheer driving pleasure with a groundbreaking approach to electric vehicle design that transforms automotive manufacturing and is powered by data to advance performance and safety. BMW's global engineering team tests the EVs in diverse driving scenarios, from long distance to extreme weather.
Using HPE GreenLake, data is processed from each vehicle on an edge computing platform, assisting the team in leveraging analytics and machine learning, engaging AI to deliver vital insights, all allowing for seamless collaboration, development, and a greater speed to market. This is what we unlock when intelligence has no limits. You drive with better performance, more confidence, more sustainably. We drive a data foundation, advanced connectivity, faster innovation. You drive with joy. We drive towards an industry transformed. Unlock adventure, collaboration, joy. HPE unlock ambition. three, two, one. The moment has arrived and so have you. You're more than a guest. You're the main character.
Surrounded by ocean, whatever you want is at your fingertips.
Carnival Cruise Line is partnering with Hewlett Packard Enterprise to ambitiously redefine the vacation experience. A floating city miles out at sea brought to life by HPE Aruba Networking with seamless shipwide connectivity from your first step on board, everything feels personal, responsive, and effortless. The crew is always on hand, quietly working behind the scenes to make you feel like a VIP, and facial recognition enhances that experience from letting you board safely, quickly, and smoothly to helping you capture your memories during the cruise. Making it all possible are two HPE-powered data centers, essential technology when you can't afford downtime this far from shore. Shipwide connectivity gives the crew instant insights on everything from weather patterns to performance and security data that anticipates needs before they even arise, so the crew can focus on those small, thoughtful details that make every guest feel truly cared for.
Together, HPE and Carnival Cruise Line are expanding your imagination and how far travel can take you. Unlock fun. HPE unlock ambition. Today, countless times all over the world, a plane will leave the runway carrying passengers across the horizon. The doors will open to new adventures, friendships, even new lives. Flight has freed us for endless possibilities. Now aviation itself can aim higher to bring the world closer with more efficiency and less impact on the environment. GE Aerospace's ambition is to decarbonize commercial flight with jet engines that enhance performance, reduce fuel consumption and greenhouse gas emissions, and take the industry in a new direction. At Oak Ridge National Laboratory, GE Aerospace engineers use an HPE AI supercomputer to run simulations at close to a quintillion calculations per second to optimize components, testing countless versions to achieve their breakthrough design.
This is what we unlock when intelligence has no limits. While you explore navigating new landscapes, embracing new cultures, discovering new perspectives, we explore aerodynamics, acoustic physics, engine performance. You explore endless horizons. We explore the future of flight. Unlock exploration, determination, curiosity. HPE. Unlock ambitions. Football, soccer, the world's game. It's more than just a sport. It's a global phenomenon that unites billions. Every roar, every chant, every goal. They are more than just moments behind. Every moment a story. Behind all the data, a fan. Now, FC Barcelona is teaming up with HPE, a world-class cloud platform networking. Solution that sets a new standard.
If you could make your way back to your seats, we're getting ready to. Get started with our Q&A. Please find your seats. Thank you.
Part of the game. Together we're building more than just a state stadium. We're building the future of football. Espai Barca. Powered by innovation, fueled by passion. When harnessed.
Let's get started. Thank you, everyone. Okay, we're not quite ready yet. Thank you. I've got to introduce people. Thank you all. As those of you who don't know, I'm Shannon Cross, Chief Strategy Officer here and obviously used to be one of a lot of you. Very happy to have everyone here. For our Q&A session. Today we're obviously welcoming Antonio, Marie. Rami, back to the stage.
Joining them we have Fidelma Russo, who's our EVP and General Manager of the Hybrid Cloud business. She's also our Chief Technology Officer. We have Neil MacDonald, who is our EVP and General Manager of the Server business. We have Maeve Culloty , who is President and CEO of HPE Financial Services. Before we begin, please do review the risk statement that was referenced by Paul at the beginning of today's meeting. Additionally, since this session is being.
Webcast, kindly state your name and company. Before posing your question, we will have mic runners, so we want to make sure everyone can hear your questions. For those of you who are joining us via the website, we do invite you to submit your questions online. We also ask that you limit yourselves to one question only. With that, let's go to Amit. He was first to raise his hand.
Thank you. Amit Daryanani, Evercore. I have to say I really appreciate. The 90-minute presentation is concise, and I appreciate it a lot. I think one of the. Things that I think folks will struggle with is the networking guide for next. Year of low to mid- single- digits. I was hoping if you could just talk a little bit in terms of what's going into that assumption because you. know from afar it looks like Juniper Networks' order momentum was very strong exiting last year. That should give you a lot of m omentum on the growth. Just talk about what are you a ssuming this low to mid- single- digit revenue expectation on the networking side. Marie, the free cash flow number for 2026?
That's two questions.
All right, I will stop.
Why don't you start with and then go to Rami for some insights.
Yeah, I mean we just started the integration of the business. We believe we have tremendous opportunity. When I think about the campus and branch, definitely we have momentum whether it was HPE Aruba or Juniper Networks. Obviously, when you think about the AI data center switching, it's going to take a little bit of time because as you go through, there is a footprint to be won which takes cycle because you have to win the technical reference. As Rami said, there is a process to become more intimate with some of these new cloud service providers and the like. Also, Rami is integrating the switches and any other assets with Fidelma in Neil's portfolio. We want to make sure we give the teams the time to do that.
We have a big sales harmonization that's taking place at the beginning of fiscal year 2026 because as you recall, we are finishing our fiscal year October 31st for HPE. Rami and team are still working all the way to October 31st and after that, there is a period of harmonization for which we have great plans. The reality is that I feel very good about the campus and branch. Routing is a cycle, we know that, and that will be a lot attached to the AI deployments, SASE and SSC, and all of that is already gaining momentum. Ultimately, the speed of that growth will also be dictated by the growth in data center network. I don't know, Rami, if you have any.
I think you covered it well. I mean, you know our business pretty well. When you're looking at our revenue growth for next year, just keep in mind that the bulk of our business is campus and branch and routing infrastructure. These are great markets, but they're not the fastest growing, right? Data center is growing very fast right now. We all know that that doesn't represent a very big part of our total business, but certainly it is a massive opportunity for us to go and pursue together. It happens to be the area where I'm most excited about revenue synergies. We did not bake revenue synergies into 2026 for a specific reason. I have a lot of work in 2026 to integrate two businesses together without disrupting any customers. I think I can do that. We were a bit cautious about expecting too much in that time frame.
I think over time, through both commercial and technical integrations, we can see revenue synergies. That's just going to take a little bit of time.
Consider prudent.
Yeah.
Let's go to Wamsi.
Thanks for handing over the mic. Thank you for doing this. Appreciate it. Wamsi Mohan, Bank of America I guess. On free cash flow, Marie, we're starting off 2026 at $1.5 billion- $2 billion, and it's going fairly significantly to $3.5 billion. Can you talk about some of the puts and takes? How much of restructuring is embedded into? Your fiscal 2026 numbers? Is there any in the fiscal 2028 numbers? You made a comment about seasonality. Could you just address if the fiscal 2026 seasonality was a comment about both revenue and earnings, or was that just an earnings comment? What maybe some bookends for that might be.
Sure. No thanks, Wamsi. Maybe I'll hit the second part of your question first. The seasonality comment was really in regard to EPS because as you think about it, the synergies themselves are going to materialize throughout 2026. It will take time and they'll be mostly back-end loaded. Revenue will follow a normal sort of seasonal path, but it'll be much more in terms of the EPS growth itself. With respect to cash flow, as you bridge the puts and takes from 2026 to 2028, I first of all start out by looking at the actual just the earnings growth itself. If you look at the sort of top line, I mean we're going to see single-digit revenue, but you're seeing more than 2x sort of growth in terms of earnings power.
You're seeing the real two companies coming together really unlocking that capability as we get into 2028. As you bridge down from that, I think as I mentioned in my prepared remarks, most of the cost of both Juniper Networks and Catalyst initiatives is really anchored in 2026 and that's why you see the cash flows sort of walk from 2026 to 2027. Then by largely by the end of 2027, most of the restructuring costs are out and we see 2028 is a pretty clean representation of what our cash flow should look like for both combined companies. Plus we've got some better working capital in there as well.
Actually, one way I like to explain the different way, if you look at the true generation of cash, it's actually close to $2.4 billion, $2.5 billion. Then you have to pay for the $600 million related to Oini and the expenses of the Juniper synergies, and obviously Catalyst. The way I think about this is more like $700 million this year because we closed the transaction and we had a lot of working capital tied to AI. Next year is think about that too on a progression to more than $3.5 billion. As we go to 2027, 2028, the expenses start going down, whether it's interest on the debt repayment, which we are committed to bring down to two times leverage in 2027.
As the expenses of the restructuring related to Catalyst and the synergies on Juniper go down, you have the extra lift as we continue to expand our operating profit, which is obviously double- digits.
Thank you.
Let's go to Aisha since you've got t he mic, and then we'll go over t o the other side.
Great, thank you. Asiya here from Citi.
I'm sorry.
That's all right. From Citigroup. As you look at, you know, there's pretty significant TAM that we're showing shared as you think about, you know, what HPE combined with Juniper, HPE's market share could be across those various segments that you talked about. If you can level set, you know, where are you now, what does that revenue growth expectations that you've outlined, what kind of share aspirations that. You have across networking and now you're. You know, refined segment of cloud and AI across that sum. Thank you.
Let's start with our traditional businesses, right? Obviously in the server business we have a large footprint, but there is profitable share growth, not share for the sake of share. That has been our strategy for a longer period of time now. As I said in my prepared remarks, it is balancing units growth or unit market share with the expansion in AUP to drive the services attached. That ultimately makes us in the traditional servers probably the most profitable business in the server category. There, that's the balance. In storage, you saw that we gained 3 points of share this past quarter. That's unheard of. Fidelma has been in the storage business much longer than I and we've never seen 3 points of share in just one quarter.
That's because we are growing seven times faster than the market and there we expect to continue to gain share. Our share combined is in the, call it, 10%- 11%. We expect to continue to grow that share. In composing branch, starting with networking for example, obviously we have a now larger footprint. Rami talked about being already 18% market share. There is a large pool of people there that are competing. Today is at least six to seven, just here in the United States and obviously outside the United States you have more players. Our goal is to gain more share also outside the United States because that's an opportunity, particularly with Chinese vendors being dislocated. With that in mind, our goal, and Rami can speak more about it, is to become the number one in Wi-Fi.
We already have market leadership and that also drives an opportunity in the campus and branch switching because every time you do this refresh from 6- 7- 8, eventually there will be a campus branch switching refresh. Now we have a complete portfolio. Through the convergence with security, obviously we have an amazing converged security offering now combining Juniper Networks assets and HPE Aruba Networking with our Silver Peak acquisition we did with SASE and SSE. That's our aspiration.
AI is hard to define what the market share is because ultimately every day, every week the market, you know, with announcements of investment, the like, it's hard to predict, but our goal is to gain share in the sovereign space where we already have a large footprint with our supercomputing, obviously in enterprise, which is our main focus through the offers between Fidelma's offerings in HPE Private Cloud AI and Neil's offer in the AI at scale, so servers, and there we have the right to play not just because our infrastructure, but because of our HPE GreenLake value proposition and the integration of the software that comes with it. Anything you want to. Okay.
Okay, Aaron.
Thank you. Aaron Rakers with Wells Fargo. Rami, I'm going to ask you a little bit more deeper about networking. If, I guess if I'm doing the pro forma numbers right from this. Last question, last quarter Juniper Networks grew like 20% some odd percent. It looks like very strong quarter for you guys. I'm curious, how would you help us understand your positioning in some of t hese cloud AI opportunities? Where exactly do you find Juniper's position today? Do you have any thoughts around, there's a lot of architectural stuff going on. Where do you guys stand as far as like co-packaged optics, scale up opportunities, etc ? Anything that you'd like to talk to? Thank you.
Yeah, certainly. This is the network for AI opportunity, basically the infrastructure that goes into data centers. There are really three distinct opportunities that are really interesting for us. The first is in data center to connect. I talked about this earlier. It's about connecting AI data centers together. Many of the big cloud builders, including hyperscalers, are building out their WANs, wide area networks, to connect data centers together. We have significant footprint here that we can leverage to upgrade 400 gig, 800 gig. The next is a direct cloud connect opportunity.
All the hyperscalers need sophisticated routing infrastructure to offer direct connect services as on-ramps to their AI cloud data centers. This is right up our alley, especially with the MX product line, because it's not just about speeds and feeds, it's about logical scale. We can have an offline discussion about what logical scale means. The third is in the data center, data center switching infrastructure. Here, as you know, not in the hyperscalers, but anything other than the hyperscalers, we have significant wins from small, medium, and large, front end and back end. The opportunity there is immense because there's massive investments that are happening there. Those are the three opportunities we have to pursue.
Think about the relevance we just gained over the last few months since the close, Neil, Fidelma, all these, Antonio, pulling me into conversations with cloud providers, whether they be Neo clouds or sovereign cloud clouds around the world. That's pretty interesting. It is just going to take us a little bit of time. As far as co-packaged optics, liquid cooling, these are absolutely going from a mode of nice to have to you will not be able to compete in the 1.6 terabit Ethernet generation without it. Our position in being able to implement these technical capabilities in our switching has gone from where everybody else is, which is experimentation, to years and years of actual production products, which is game changing.
Okay, Samik
Thank you Samik from JPMorgan . Maybe if I can change gears and ask you about the AI compute business specifically where you're outlining the focus on enterprise and sovereign customers in the backdrop where we've seen a significant amount of announcements from Tier 2 CSPs. How are you trying to balance what guardrails are you putting on that business? In terms of businesses that you go f or margins on that AI infrastructure server business. To sort of balance profitability and growth, how should we think about, as you sort of think, how should we think about enterprise and sovereign mix tracking from here on? Should we expect that mix to increase or sort of stabilize? At this sort of 50% level that you've outlined for the last quarter?
Do you want to?
When you think about the structure of the market, there's a sizable volume involved in big model builder and service. Provider buildouts and there's a very r apidly growing set of opportunities in sovereign, which also includes some service provider space, and in Enterprise. In Q3 we saw 250% growth in our bookings in the sovereign space, and we're very excited about the momentum that we've had in recent months. As Antonio mentioned, we've commissioned the U.K.'s sovereign AI infrastructure that we delivered. We've commissioned the French [Foreign language] and their infrastructure for AI, which is the largest classified AI system in Europe, as just two examples, and there are many others in which we're engaged. Part of what enables us to create value there is the ability to shorten the time to deployment. With the work we do around modular data centers with the integrated direct liquid cooling, we can deliver these systems to production much faster than these entities can. That's been a driver for us. We will continue to be very selective in model builder and hyperscaler opportunities in this space, being careful to ensure that our requirements around working capital and profitability are balanced with the need to participate in that space for the halo that it creates in the enterprise market. We are also seeing momentum and growth in where our reach and our services. Delivery capability creates more value that sometimes isn't needed by the big model builders and the hyperscalers.
Okay, Erik. Oh. It's a race.
Hey guys. Than k you. Erik Woodring, Morgan Stanley. Rami, I wanted to build on Amit's question, which was, you know, you started the presentation about kind of 9% CAGR in the networking market and how core it is to kind of the Newfound HPE over that same period y ou're guiding your own networking business to 5%- 7% growth, inherently implying t hat you're losing share. Can you maybe help us s quare together that dynamic of kind of guiding to losing share versus your clear k ind of optimism around what HPE a c ombined Juniper Networks and Aruba, each networking business could do?
Yeah, just again, you keep in mind you've got to look at the mix of our business, right? The mix of our business is not identical to the mix of the total TAM between data center, campus, and branch and routing infrastructure. The bulk of our business is in routing infrastructure and campus and branch. These two market opportunities are growing in the low to mid single digit range. You are not going to see a total growth rate for HPE networking that's going to look like the total growth rate of the total networking TAM that has a totally different mix shift. Just keep that in mind. To be a bit more specific, we're going to take share in campus and branch, we're going to take share in routing infrastructure. I believe we can even take share in SASE and security.
We're being a bit prudent right now in data center as we go through the integration of our two businesses. There are a lot of dynamics happening in the data center space between white box switching, Celestica coming into the mix, NVIDIA coming in with their products. There are a lot of dynamics, but at the same time I will reiterate if there's one area I am most excited about from a revenue synergy standpoint, just because of the relevance that HPE brings to the table. The ability to leverage liquid cooling, direct liquid cooling technology, the ability to open doors for us internationally where Juniper Networks had very little presence, the ability to give us more relevance with the AMDs and NVIDIAs and the other GPU providers of the world. That is something that's very exciting. I'm, you know, we haven't baked that into our numbers.
Yeah, look, Erik, different way to think about it. We will grow faster than the market in campus and branch and routing. We believe we can do that the same in the security, in the convergence space. We have seen that in our own numbers. When it comes down to SASE and all of that, and then in data center networking, right, obviously we will see the benefits in the enterprise data center because the work Rami is doing with his colleagues. In the AI space, just by winning one customer at some point, that simplified the whole story. That may take a little bit of time because he needs to go build that intimacy now through the door openers that we are driving with the rest of the team, and there is continue to sell to the ones we have and then get new logos.
Rami is right, there is a lot of transitions that happen next year in technology. We want to be there first with the 1.6 Tb. That will give us a footprint, and the sales cycle is a little bit longer, and after that, time to revenue is a little bit longer because it's a little bit lumpier as well. Once you win one or two customers, the whole thing becomes bigger, and therefore we have been prudent making numbers that we don't have yet. The whole story, lineup from an integration technology, and then go to market, which we are building as we speak.
Okay, let's go to Simon.
Thank you very much. Simon Leopold with Raymond James . First, just a quick clarification if I might t he growth rate for fiscal 2026, if y ou could clarify the baseline given that y ou acquired Juniper Networks midway through the year. The thing I wanted to ask. What was your vision of how enterprise adoption of AI initiatives affects spending in? That how much will your customers employ on the public cloud vs investment in their own data centers, which is really your served market. How do you see that playing out? What will they buy, and when will they buy it? Thank you.
Maybe you want to answer the baseline.
Yeah, it's. As I said, the 2026 guide is low to mid and that's off a pro forma basis and including Juniper. We moved to 5%- 7% in the long-term guide. This is all pro forma. You should see that in the 8K that came out in terms of the pro forma financials.
Yeah, we'll have the 8-K out. It should be out now. You'll be able to apologize, rebuild your models.
As you look at the reporting, it's going to be much higher because, remember, we have four months of Juniper Networks in our numbers this year and 12 months next year. So. It is going to be very high double- digits by reported standards. On the private cloud versus the public cloud, maybe Fidelma, you want to take that question about how enterprises are using a combination of both.
As we've gone through the last maybe two years with really engaging with enterprises of all sizes, what we've started to see is within very, very large enterprises, they look like CSP at the top end, but within their IT organizations where they don't have the skills and where they're really trying to drive productivity across different functions like support help desks, that's where our turnkey system like HPE Private Cloud AI is really beginning to play. We're seeing growth quarter over quarter and it's beginning to accelerate. That gives us, you know, the market in the enterprise is a lot slower. We're also seeing people starting to experiment on the public cloud and then when they go to deploy within the enterprise, they're making the choice because of cost, security, compliance to move it on prem. We're starting to see the acceleration of that motion through enterprise customers across the board.
If you think about certain segments of enterprise, while their technical requirements look a lot like the service providers, as Fidelma quite rightly just. Said that's not their core business. There is a lot of opportunity for us to deliver infrastructure at scale that looks like service provider scale, but with a much richer set of capabilities for delivery and service. We've been doing that with customers. Around the world all the way through FY 2025,
And it's true in financial services. All right, I don't know.
We've got Tim over there.
Yes, we'll be back.
Thank you. Tim Long at Barclays over here. Two parter i f I could, maybe we could touch on HPE GreenLake a little bit. Curious on two aspects. Number one, it's obviously been growing nicely over the years. What does the addition of Juniper mean? To the pipeline for HPE GreenLake? Number two, kind of related to that last question, as enterprises evolve with their AI journey and start to bring more on premise, do you see that potentially as a catalyst to ignite some higher growth in HPE GreenLake?
Fidelma?
Thank you for the question and the acknowledgment that it's had nice growth over the years. Rami and I see tremendous opportunity for Juniper within GreenLake. In fact, sometimes when we deploy a GreenLake solution, we actually already have Juniper switches. In deploying GreenLake.
Very strong a lternative to the current incumbency. That's an example. Also, in the next few months we're going to integrate Juniper Mist as a launching platform with HPE GreenLake as Rami drives the convergence as he spoke about it.
Okay, Lou, next.
Thank you, Louis Misciosci a, Daiwa Capital Markets America. Maybe this is a good summary question. You explained networking really well. What's prohibiting growth there? With such great assets that you have—AI, private cloud, on-prem, server modernization—I guess what's holding back possibly having faster growth? Is it really just what you had? just mentioned earlier that enterprise companies just aren't moving to AI fast enough, that they're stuck in POCs and just haven't really deployed that. If that is the case, what visibility do you have? Does it start to hit late first half 2026, second half 2026 calendar year? Anything to help us understand you guys getting to faster growth would definitely be helpful. Thank you.
I can start. Lou, I think, as I think about the segmentation of the market for AI, obviously Neil spoke about the service provided to Simon's question. The model builders, they are fewer, their customers think about tens. That drives millions of GPUs now. We'll see when all of that gets built. You have enterprises like us, which I hope you got that we are very aggressive deploying AI. As you can see, Marie is very enthusiastic about that in her own organization. They are talking about a very smaller number of AUP, call it, because ultimately it's not in the hundreds of millions or billions. You're talking about single millions. You have to pile up all these transactions to start making a dent in the overall numbers.
The good news is that since we started with HPE Private Cloud AI and the enterprise focus, we have been growing every single quarter. One of the things I mentioned in my remarks is that we already have a $500 million pipeline in the HPE Private Cloud AI. That is something we are really focused on converting to, and that conversion happens through the POCs directly with our customers and through our channel partners because we are enabling all the channel partners with that infrastructure so that they can bring the customers to their offices. What they are doing is proof of concepts when they decide to do there versus the public cloud. Eventually, once they are ready to go because they figured out the use case, the return on investment and all that, we actually help deploy that with advanced professional services and our support services.
In addition to the fact that one area that Maeve is triple focus is financing AI in enterprise. Sometimes some new clouds, sometimes some sovereign customers, sometimes. That's where our focus is. We expect that to continue to grow very nicely in 2026 as we go forward.
Marie, you want to talk just a little bit, you've been dealing with what she's been doing in AI with Deloitte and others. You know what you've heard from some. Of the CFOs and other companies you've been engaged with.
Yeah, look, I'd say that folks definitely, having spoken with a lot of CFOs out there, are in early stages of trying to understand how enterprise AI will really transform the business. What I would say is, you know, we've seen some pretty amazing, dramatic transformations that have happened, but we're probably at the early innings. That's why you also see in our margin guide we actually increased our long-term margin rate from 26%- 27%-28% because you start to see enterprise and sovereign become more significant in those years. It's going to take time. We're in the early innings.
Okay, David?
Great. Thank you, David Vogt, UBS. Maybe for Antonio and Rami. Rami, in your prepared remarks you talked about meeting your customers' needs with both the Aruba portfolio and the Juniper portfolio. Can you expand on sort of the go to market for driving those separate? Businesses discreetly and then ultimately from a s ynergy perspective, how do we think about it t hat in the context of your longer t erm guide in terms of that part of the market, not on the data center side, but o n the campus and branch side?
First, there are two great platforms and honestly, now that I've had an opportunity to look under the covers and the Aruba side, I'm really impressed. It's a bit of an embarrassment of riches right now in terms of what we have under our portfolio. Second, they both are AIOps platforms, but they both have unique strengths and capabilities. Mist is a public cloud only AIOps platform. As great as Mist was in AIOps capabilities, we were shut out of any opportunities that required different deployment models like private cloud or virtual private cloud networking or even on premises. When I looked at Aruba, they've actually made more progress in areas like security integration, agentic AI capabilities. They really have unique strengths. The second thing that's really important to understand about these platforms is that these are not monolithic code bases.
They were developed using a micro-services-based architecture and it is actually quite straightforward to take micro-services from one and apply it to the other. That's the plan. We can keep these platforms and their unique deployment capabilities in place and start to cross-pollinate. Mist has done amazing work in AIOps space. I can take that as a micro-service and apply it to Aruba. Aruba has done great work in security integration that can be a micro-service that can go and apply to Mist. In so doing, what have I done? I've accelerated the overall pace of innovation on both platforms. I've made my engineering more efficient because I don't need to develop something twice. I can develop it once and deploy it twice. That's the grand plan that we're working on right now.
I think the industry is sort of a little bit tainted right now because our peers in the industry have taken years to try to figure out how to integrate portfolios and we're showing our customers, much to their amazement, that we can do it in a much, much less period of time. Watch the space.
And I'll just a dd on the guide in terms of the long-term guide, the portfolio simplification and sort of R&D project overlap that we talked about earlier is included in the $800 million of synergies as well.
Yes, and one of the key differentiations also why he can drive that convergence is HPE GreenLake and all the cloud, the common cloud platform and the common cloud services sitting underneath that Fidelma and team provides to everybody here on the stage. As we go through that journey that Rami just described, cross-pollinating and doing things once that convergence happened very naturally, and the user interface will become one at some point, but no customer gets left behind. They are very excited about that, and then they get more faster.
Okay, and we'll take our last question here, Mike.
Thank you. Mike Ng from Goldman Sachs. I wanted to just ask one on networking and just have a quick financial follow up on automated WAN. I think when Juniper Networks was a standalone company that was more of a flattish growing category. Could you just bridge what's gotten better so that we're growing at that 5% CAGR now? Is it mostly DCI, Direct Cloud on-ramps, or are there other elements that are getting better as well? Could you just talk a little bit about the networking margin, going from low 20s next year to 25%- 28%? Is there a mix element there, with DC switching contributing more, or is it just cost saves? Thank you.
I'm happy to address the WAN question, and Marie, maybe you can talk about the margins. Take the margin question. On the WAN side first, Juniper actually over the last year or so, even during this closed period, has seen really good tailwinds in WAN primarily because of what you just described. Data center interconnect cloud providers need wide area networking to connect AI data centers together, and they need it for the on ramps to the data center. We see we have some really great solutions in that space. Another trend that I expect will get better is routing for enterprises. Many large enterprises actually require routing at the edge or even run their own wide area networks. We at Juniper were always limited by just the scale we had in go to market.
Yeah.
All of a sudden, you know, we've got a much, much bigger enterprise salesforce that can sell not just compute and storage, but we're going to teach them how to sell routing as well.
Yeah, I'll just add that on the margins themselves, in the long term guide we expect to get to 25%- 28%, obviously coming off the low to mid in 2026. That's exactly right. You answered the question very well. It's a combination of the cost synergies, which will take some time, and then the scaling of revenue growth and the benefit you see in terms of just the model itself in the outer years. I think we're really, really pleased with those long term targets.
Great. I think we've concluded Q and A. Antonio, I think you had some closing remarks.
Thank you for joining. I'm going to stand up because I want to address the webcast too, so you guys can stand. Thank you. Thank you. I know we are going to spend some more time here together, but I want to leave you today feeling confident that HPE is more than prepared for what is next. In fact, we are actively shaping our future for our customers. We will continue to push the limit of innovation. One of the areas I'm incredibly proud as a company we never stop innovating and if you think about the portfolio that we have today, it is something that we could only dream seven and a half years ago. I'm very proud of what I have done there. For our partners, we are opening the door to unmatched opportunities for collaboration, scale, and assured success.
We do a lot of business with our partners and they are coming along with us. For our team members, which are the core of our culture, we are committed to create opportunities for career-defining work with a culture that's consistent, recognized both internally and externally. For our shareholders, which many are here, we are offering a very compelling investment opportunity, an opportunity built on that profitable growth and strong returns, including as we announced today, growing our annual dividend and increasing our share repurchase as we execute the strategy we just discussed. I believe the future is here and I believe HPE is leading the way and together we will redefine what is possible in technology to deliver that lasting value for all our shareholders. Thank you for joining us today on the webcast and thank you for joining here in person.