HPE wrapped up fiscal 2025 with a sharp acceleration in AI infrastructure and networking demand, booking $9.7 billion in Q4 revenue—up 14% year-over-year—as the company began to absorb the Juniper Networks acquisition into a combined portfolio aimed squarely at next-generation AI data centers. Networking revenue jumped 150% to $2.8 billion, driven by strong adoption of HPE Juniper WAN, data center switching, Ultra Ethernet–ready platforms, and new integrations with NVIDIA’s Spectrum-X fabric. HPE closed the year with $6.8 billion in cumulative AI server orders and a record AI backlog that continues to grow with sovereign and enterprise projects .
AI infrastructure remained the performance anchor of the Server and Cloud & AI segments, even as traditional server revenue slipped 5% in the quarter. HPE’s AI order book—now exceeding $11.5 billion cumulatively across FY23–FY25—and its new data-pipeline products, liquid-cooled switches, and rack-scale supercomputing systems helped lift non-GAAP operating profit to $1.2 billion in Q4, up 52% sequentially. Q4 free cash flow reached $1.9 billion, powered by AI system deliveries and continued GreenLake expansion. ARR climbed to $3.2 billion, up 62% from a year ago, with software and services contributing over 80% of the total.
The integration of Juniper contributed four months of revenue this year and helped position HPE as a top-tier AI networking supplier entering 2026. HPE reaffirmed its FY26 revenue growth outlook of 17–22%, supported by mid-single-digit pro-forma networking growth, double-digit AI system shipments, and expanded NVIDIA and AMD ecosystem partnerships. The company expects most of its AI backlog to convert in the second half of FY26 and into FY27 .
Key points
• Q4 revenue: $9.7 billion, up 14% YoY; Q4 non-GAAP gross margin: 36.4%, up 650 bps sequentially.
• Networking: $2.8 billion revenue, up 150% YoY; 23% operating margin; ARR up 200% YoY.
• Server: $4.5 billion revenue, down 5% YoY; AI orders reached $6.8 billion cumulative since 2023, with >60% from sovereign and enterprise customers.
• Cloud & AI: Combined GreenLake ARR reached $3.2 billion; more than 46,000 GreenLake customers.
• AI Systems: Q4 AI backlog reached a record high, with most revenue recognition expected in 2H FY26 and beyond.
• Networking for AI: Juniper QFX 5250 liquid-cooled switch, PTX/MX integration with NVIDIA Spectrum-X, and AMD “Helios” rack-scale integration.
• Cash: Q4 free cash flow reached $1.9 billion; FY25 FCF totaled $986 million, above full-year guidance.
• FY26 guidance: Revenue $9.0–$9.4 billion for Q1; full-year non-GAAP EPS $2.25–$2.45; FCF $1.7–$2.0 billion.
• Dividend: $0.1425 per share, payable Jan 16, 2026.
“HPE finished a transformative year with a strong fourth quarter of profitable growth and disciplined execution,” said Antonio Neri, president and CEO of HPE. “We completed the Juniper Networks acquisition, scaled our AI and cloud businesses, and accelerated innovation across our portfolio.”
Additional Insights from the Q4 FY25 Earnings Call
• Late-quarter order surge: HPE reported a sharp acceleration of orders across networking, servers, and storage in the final weeks of Q4, driven partly by customer budget deadlines and advanced notice of DRAM-related price increases taking effect in November.
• Pricing actions already implemented: HPE confirmed it raised server pricing in November to offset DRAM cost inflation—describing expected DRAM and NAND increases in early 2026 as similar to early-COVID supply allocation cycles.
• AI server shipment pushouts: Several large AI system deals slipped from Q4 into Q1 due to customer data-center readiness issues and slower U.S. federal procurement reactivation.
• AI pipeline far exceeds backlog: HPE emphasized that its AI systems pipeline remains “multiples” of backlog, with sovereign customers placing larger, longer-lead-time orders tied to NVIDIA Vera/Rubin-class platforms.
• AI conversion now heavily back-loaded: Due to sovereign project timelines, infrastructure buildouts, and long qualification cycles, HPE expects the majority of AI revenue to land in 2H FY26 and into FY27.
• Networking profitability milestone: Juniper delivered its highest operating margin in eight years during Q4, driven by early synergy capture and strong routing/DCI demand.
• Sales-force integration timing: HPE confirmed that the Aruba + Juniper networking sales teams merge on January 1, with a unified compensation plan launching immediately—cited as a major FY26 execution milestone.
• AI networking traction: Juniper MX and PTX platforms were highlighted as key wins for AI edge-on-ramp and long-haul DCI; HPE expects significant backlog conversion later in FY26 given Juniper’s historically back-end-loaded revenue pattern.
• Traditional server AUP strength: Gen11 and Gen12 servers now make up 98% of traditional server revenue, significantly lifting average selling prices as customers modernize to reduce power usage and core footprints.
• Demand elasticity management: HPE disclosed it is actively using “demand shaping” and dynamic configuration strategies to navigate rising DRAM costs without eroding margins.
• Hybrid Cloud margin strategy: HPE reiterated an intentional reduction of low-margin non-IP business, contributing to near-term revenue pressure but improving long-term profitability.
• Financial Services strength: The financing arm posted its highest ROE in more than five years (21%), with a steady 0.5% loss ratio and more favorable leasing mix.
• Cash conversion improvement: HPE reduced its cash conversion cycle to 30 days, helped by exceptionally strong Juniper receivables collections.
• Inventory drawdown: Company inventory fell 19% YoY and 11% sequentially, tied to AI server shipments and tighter supply-chain controls.
• H3C sale fully baked into FY26 OI&E: Management confirmed that the $1.4B H3C divestiture and related financial impacts were already incorporated in forward guidance.
• Back-half weighted EPS: HPE expects 53% of FY26 non-GAAP EPS to materialize in the second half due to AI shipment phasing, networking seasonality, and commodity-cost timing.
🌐 Analysis
HPE enters FY26 as a consolidated AI infrastructure supplier with a full stack spanning servers, AI fabrics, storage pipelines, and cloud-managed operations. The Juniper integration materially shifts HPE’s networking profile, giving it a stronger position against Cisco, Arista, and NVIDIA’s expanding Spectrum-X ecosystem. The company’s sovereign AI pipeline continues to stand out relative to peers, supporting long-term backlog visibility as countries build domestic compute capacity.
