Hewlett Packard Enterprise (HPE.N): HPE: Leaning Into Networking; Reaffirming Robust Demand Outlook

研报英文原文证据摘录

Hewlett Packard Enterprise (HPE.N): HPE: Leaning Into Networking; Reaffirming Robust Demand Outlook

Hewlett Packard Enterprise (HPE.N)

17 June 2026 Citi Research

accelerated significantly over the last six months as customers move from

evaluation toward deployment, particularly as they modernize infrastructure to

support on-prem AI workloads, reduce footprint, lower cooling and power

requirements, and improve performance density. Management suggested this is

not simply a refresh cycle, but rather a broader infrastructure modernization cycle

being catalyzed by AI.

Enterprise customers continue to absorb elevated hardware pricing without

slowing projects, as the need to modernize infrastructure appears to be

outweighing budget pressure. Management said customers are still trying to

extend where they can, but the urgency to support AI on-prem is increasingly

forcing upgrades, particularly given the associated space, cooling, and power

requirements. HPE also emphasized that Financial Services and GreenLake are

helping customers manage the transition by freeing up capital and enabling more

OpEx-oriented consumption models, while management expects the cost curve to

stabilize somewhat in FY27 but remain elevated.

AI demand is increasingly centered around inferencing and on-prem enterprise

deployments, though management still views the market as early. Management

said the current training-to-inferencing compute mix is still roughly 70/30 but

expects that to shift materially over time and believes at least 80% of AI demand by

2030 will be inference-related. They also highlighted that once customers attach

proprietary data to models, many will increasingly want those workloads under

their own control, particularly in regulated industries, which supports the case for