Hewlett Packard Enterprise reported 40% revenue growth to $10.7 billion in its fiscal Q2 2026, driven by a surge in AI server demand that lifted its AI systems backlog to $5.9 billion.

HPE's server revenue rose 33% year over year to $5.5 billion in fiscal Q2 2026 (ended April 30, 2026), with the company receiving $1.8 billion in new AI systems orders during the quarter. However, reported top-line growth of 40% was significantly boosted by HPE's acquisition of Juniper Networks, which closed last July; networking revenue surged 148% year over year but grew only about 10% when stripped of the deal.

AI demand is broadening beyond model-training clusters: inference and agentic workloads (AI systems that make autonomous decisions) are now driving traditional server sales alongside purpose-built AI systems. HPE unveiled a new ProLiant server built around Nvidia's Vera CPU at COMPUTEX, aimed at these workloads.

Supply constraints, particularly in memory, are capping how quickly HPE can convert orders into revenue, and the company expects elevated costs to linger into 2027. HPE's cloud and AI segment posted a 12.4% operating margin in the quarter, nearly double the year-ago 6.6%, but well below its networking segment's 21.6% margin, reflecting the low-margin nature of selling hardware tied to others' accelerators.