Navigating the AI Infrastructure Boom with a High-Margin Networking Pivot

Hewlett Packard Enterprise (NYSE: HPE) is a global edge-to-cloud technology company formed in 2015 from the split of Hewlett-Packard. The company designs, manufactures, and sells servers, storage, networking, and software solutions, with a strategic emphasis on hybrid cloud, AI, and high-performance computing. Revenue is generated through three primary segments: Compute (servers, approximately 40% of revenue), Networking (Juniper-integrated switching and routing, approximately 25%), and Storage/Intelligent Edge (approximately 20%), with the remainder from financial services and consulting. Customers span large enterprises, public sector institutions, research organizations, and service providers across 170+ countries. HPE employs approximately 62,000 people worldwide and serves over 65,000 direct enterprise accounts, with an additional 100,000+ through channel partners. The company's scale is substantial, with FY2025 revenue of $32.1B, and its strategic pivot from commodity hardware toward software-defined and AI-centric solutions is reshaping its revenue quality.

• Near-Term (FY2026-2027): AI server revenue is projected to grow 30-40% annually as enterprise inference workloads (not just training) begin to scale. The Juniper integration contributes a full year of networking revenue, adding approximately $5.5B in FY2026 sales. Management guides to 8-10% total revenue growth in FY2026, accelerating to 12%+ in FY2027 as AI-optimized networking (the "AI fabric" for GPU clusters) becomes a major attach opportunity.
• Medium-Term (FY2028-2030): The proliferation of sovereign AI and on-premises enterprise AI deployments (driven by data residency and security concerns) plays directly into HPE's strengths. The company targets $50B+ in annual revenue by FY2029, with a mix shift toward recurring software and services that should lift operating margins from the current 7% toward 12-14%. Additionally, the emerging "AI-as-utility" managed service offering, where HPE operates AI infrastructure on behalf of customers, represents a potential $5B+ new revenue stream by FY2030.

Source: Company filings; FY2024A-FY2025A actuals, FY2026E-FY2027E analyst estimates.

Revenue growth is accelerating from 5% in FY2024 to a projected 15% in FY2026, driven by the Juniper acquisition and AI server demand. Gross margin expansion of 170 basis points from FY2025 to FY2026E reflects the higher-margin networking mix (Juniper products carry ~55-60% gross margins versus ~12% for AI servers) as well as improved supply-chain efficiency in storage. Operating leverage flows through to EPS, with non-GAAP EPS projected to grow 25% in FY2026, outpacing revenue growth. Free cash flow conversion is strong at 94% of net income, supported by a reduced cash tax rate following the Juniper purchase and working capital improvements.

The global IT infrastructure market is valued at approximately $250B annually, with the AI infrastructure subset (servers, networking, storage for AI workloads) growing at a 35%+ CAGR and expected to reach $150B by 2028. HPE competes in three distinct arenas: (1) AI servers and HPC, (2) enterprise networking, and (3) hybrid cloud infrastructure. In AI servers, HPE holds an estimated 12-15% market share, behind Dell Technologies (25%) and Supermicro (18%), but ahead of Lenovo. In enterprise networking, the Juniper acquisition vaulted HPE to the #2 position with approximately 15% share, behind Cisco (40%) but ahead of Arista Networks (8%) in the campus/enterprise segment. Key comparable companies include:

• Dell Technologies (DELL): Similar AI server exposure but lacks HPE's networking depth; trades at a lower multiple due to higher margin volatility.
• Cisco Systems (CSCO): The networking incumbent facing share loss to HPE/Juniper in AI-driven data center switching.
• Arista Networks (ANET): The high-growth pure-play in data center networking, growing at 25%+ but with a premium valuation reflecting its AI fabric leadership.
• Super Micro Computer (SMCI): A fast-moving AI server competitor with higher growth but lower profitability and governance concerns.

HPE's fair value is best assessed through a blend of discounted cash flow (DCF) and comparable-company analysis. Using a DCF model with FY2026E free cash flow of $3.5B growing at 12% for five years, then 4% terminal growth, and a 10% weighted average cost of capital (reflecting a beta of 1.44 and a net debt position), we derive an intrinsic value of approximately $58 per share. This implies the market is currently pricing in roughly 85% of our base-case cash flow projections, leaving room for upside if AI networking attach rates exceed expectations.

HPE trades at a discount to Arista on every metric, reflecting its lower growth, but at a premium to Dell on P/E, justified by the superior margin trajectory from networking. The 19.5x forward P/E is reasonable given projected EPS growth of 25% in FY2026, yielding a PEG ratio of 0.78, which suggests the stock is undervalued relative to its growth rate.

• AI Margin Compression: The AI server business operates on thin margins (8-10%) due to GPU cost pass-through. If AI systems grow faster than networking and storage, consolidated margins could stagnate or decline, undermining the investment thesis of margin expansion. Management must balance volume growth against profitability.
• Juniper Integration Execution: Post-merger integration carries inherent risks, including customer attrition (particularly in the telecom service provider segment where Juniper had strong relationships), sales force disruption, and potential product roadmap conflicts between Juniper's and HPE's existing networking lines. Any significant integration stumble could impair the expected $450M in annualized cost synergies.
• AI Demand Cyclicality: A meaningful portion of AI infrastructure spending is concentrated among a small number of hyperscalers and well-funded AI startups. A slowdown in AI investment or a shift toward cloud-only deployment models could sharply reduce HPE's AI server orders, exposing the cyclicality beneath the growth narrative.
• Competitive Pressure from NVIDIA and Cisco: NVIDIA's move into full-stack AI infrastructure (including networking via its Spectrum-X platform) threatens to disintermediate HPE in AI data centers. Simultaneously, Cisco's partnership with NVIDIA for AI fabrics could blunt HPE's networking gains. If HPE becomes a "dumb pipe" for GPU clusters, pricing power will erode.
• Geopolitical and Supply Chain Exposure: HPE sources key components (GPUs, memory, advanced chips) from a concentrated supplier base, primarily in Taiwan and South Korea. Cross-strait tensions or export controls could disrupt supply, while the company's significant sovereign AI contracts in the Middle East create geopolitical risk if sanctions or diplomatic shifts occur.