HPE Discover 2026: autonomous networking, AI agents, HPC

HPE Discover 2026 runs June 15-18 at the Venetian Conference Center in Las Vegas. CEO Antonio Neri will deliver the keynote at 9:00 a.m. PST on June 16. Official product announcements are expected after his speech, and executives and partners will use sessions across the four-day event to expand on technical and product details.

Networking is a major focus. Nearly a year after HPE acquired Juniper Networks, HPE is integrating Juniper technology into its Aruba networking portfolio. The company launched an agent-driven AIOps networking product in May that it describes as moving from concept to production. Rami Rahim, HPE’s EVP and head of networking who led Juniper before the acquisition, is scheduled to recap integration progress and the “self-driving networks” concept first shown at Discover Barcelona.

HPE has cited early customer deployments of autonomous networking across government and enterprise customers, including examples involving the UK Ministry of Justice. The company says the tools automate network actions to reduce operational tasks for customer teams.

Supercomputing will appear in sessions rather than as a headline product reveal. HPE supplies hardware for a substantial share of the world’s fastest systems: six of the top 10 machines on the November 2025 Top500 list ran on HPE Cray systems. Scott Atchley, chief technology officer at Oak Ridge National Laboratory, is scheduled to speak on Tuesday afternoon. HPE will reference Cray’s history as it marks 75 years since the first Cray supercomputer and will outline current installations and its roadmap for high-performance computing customers.

AI agents and the GreenLake private cloud platform will feature across conference tracks. HPE positions GreenLake to host AI services and agent capabilities in on-premises and hybrid cloud environments. The company shifted its positioning from “edge-to-cloud” to “edge-to-cloud-to-AI” and introduced an AMD Helios rack in 2025 aimed at cloud service providers.

Speakers are expected to address the operational and financial implications of running agent workloads, including token consumption and cost management for operator and large language model agents at scale.

Nvidia CEO Jensen Huang is not listed among speakers; a prerecorded message remains possible. HPE staff will be on site throughout the conference, including the Partner Growth Summit on Monday, to provide briefings after Neri’s keynote and during the week.