HPE Discover Las Vegas 2026: Networking Becomes the Strategy

An analyst brief synthesizing coverage from Six Five Media, Futurum Group, and Moor Insights & Strategy

‍

TL;DR: HPE Discover 2026 in Five Points

‍

• Event: June 15–18 at The Venetian in Las Vegas — HPE's first Discover since closing its $14 billion Juniper Networks acquisition in mid-2025 (Futurum Group).
• Headline: CEO Antonio Neri repositioned networking from a foundational technology into the strategic layer underpinning enterprise AI (Six Five Media).
• Financials: Fiscal Q2 2026 revenue $10.7 billion (up 40%), non-GAAP EPS $0.79 (up 108%), cumulative AI bookings $16.4 billion (Moor Insights & Strategy).
• Portfolio: Juniper, Aruba, Morpheus, OpsRamp, Zerto, GreenLake Intelligence, and Private Cloud AI now form a unified full-stack platform (Futurum Group).
• Analyst takeaway: Six Five Media's on-site roundtable concluded Discover 2026 was less about unveiling a new vision than it was about proving HPE could execute on one already laid out (Six Five Media).

‍

What was HPE Discover 2026?

HPE Discover 2026 wasn't simply a showcase of new products. It was the company's first opportunity to demonstrate how nearly a year of integration work following the Juniper acquisition had reshaped its strategy.

The event brought together customers, partners, and industry leaders across more than 225 breakout sessions, with over 80 customer speakers taking the stage. Six Five Media covered the conference live with analysts Daniel Newman, Patrick Moorhead, David Nicholson, Matt Kimball, Fernando Montenegro, Tom Hollingsworth, and Alastair Cooke.

Across keynotes from CEO Antonio Neri, Networking President Rami Rahim, and CTO Fidelma Russo, one theme consistently emerged: HPE is no longer presenting networking, compute, storage, software, and AI as separate businesses. Instead, the company is positioning them as parts of a single enterprise stack designed to support AI at scale.

Instead of focusing on individual technologies, HPE used Discover to illustrate how its growing portfolio of silicon, networking, infrastructure operations, and governance forms a cohesive platform for running enterprise AI (Futurum Group).
‍

Start With the Network: HPE's New Center of Gravity

The defining announcement at HPE Discover wasn't a product launch. It was a shift in perspective. Antonio Neri opened his keynote with a message few expected from a company historically associated with servers and storage: start with the network. It placed networking at the center of HPE's AI story rather than treating it as another infrastructure layer (Six Five Media).

Matt Kimball argued this shift represents a fundamental evolution in how HPE should now be viewed.

Matt Kimball, Moor Insights & Strategy
"HPE walked into Discover with the most coherent networking and operations strategy I've seen from the company. Networking is the catalyst and the loudest proof, but the operating layer underneath is the actual bet." (Moor Insights & Strategy)

HPE now presents itself as a networking company that also delivers compute, storage, and hybrid cloud, with Neri framing Cisco over traditional server competitors as the benchmark HPE intends to challenge. The financials back it up: fiscal Q2 2026 networking revenue reached $2.7 billion, up 10%, with gross margins expanding more than 800 basis points (Moor Insights & Strategy). Rami Rahim told Daniel Newman the Juniper integration is running "ahead of schedule," pointing to the new Helios rack-scale Ethernet fabric co-developed with AMD — the first open Ethernet solution for scale-up GPU connectivity (Six Five Media).

‍

Key networking announcements at HPE Discover 2026

• QFX5252 — Liquid-cooled scale-up switch for AMD Helios racks, supporting UALink over Ethernet as an open alternative to NVLink (Futurum Group)
• QFX5250 — Scale-out switch for AI Ethernet fabrics (Futurum Group)
• QFX5140 — Inference-cluster and edge AI switch targeting, something Matt Kimball called the largest enterprise AI opportunity (Moor Insights & Strategy)
• MX301 inference-edge router and PTX12000 DCI router (Futurum Group)
• Networking Data Center Director (formerly Apstra), unifying the QFX portfolio under a common AI datacenter platform (Moor Insights & Strategy)

Together, these updates form an end-to-end platform spanning AI training, inference, campus, edge, and data center.

The Self-driving Network Moves from Vision to Reality

Autonomous networking has long been more aspiration than reality. Marvis, the AI-native operations platform that joined HPE through Juniper, now runs inside Aruba Central, extending AI-driven monitoring and automated remediation across wired, wireless, and SD-WAN. Aruba CX campus switches can be administered through Mist, and Mist Data Center Assurance integrates with HPE Compute Ops Management and GreenLake, putting network, compute, and infrastructure operations on a common control plane for the first time (Moor Insights & Strategy).

Tom Hollingsworth noted this integration marks one of the most significant outcomes of the Juniper acquisition.

Tom Hollingsworth, Six Five Media / Futurum Group
"Networking has become foundational to enterprise AI. The integration of Aruba, Silver Peak, and Juniper has gone very well—it's now clearly 'HPE Networking,' evolving into a core AI infrastructure layer rather than supporting infrastructure." (Six Five Media).

Enterprise AI workloads place new demands on infrastructure, requiring predictable latency, continuous observability, and automated operations across distributed environments. In that context, networking is no longer simply about moving packets. It has become the control layer that enables AI systems to operate reliably at scale.

Discover 2026 reinforced HPE's broader argument that the future of enterprise infrastructure won't be managed through separate networking, compute, and storage teams using disconnected tools. Instead, it will increasingly be orchestrated through a unified, AI-assisted operations platform where the network serves as both the connective tissue and the primary source of operational intelligence.
‍

AI Infrastructure: Building the Enterprise AI Stack

Rather than introducing a new AI strategy, HPE expanded the platform it has been steadily assembling. Private Cloud AI now scales inference to 256 GPUs with NVIDIA KAI, Run:ai, a shared KV cache, and the Alletra MP X10000 data layer, with built-in agentic governance and, HPE says, “time-to-value 7 to 12 months sooner than DIY, without egress or per-token fees.” The AI Factory added Blackwell GPUs, Spectrum-X Ethernet, and BlueField-3, with Vera Rubin NVL72 on the roadmap (Futurum Group).

Patrick Moorhead sees this progress as aligning closely with where HPE has historically been strongest.

Patrick Moorhead, Moor Insights & Strategy
"HPE arrived in its best shape in years. The $14 billion Juniper deal is paying off—networking is HPE's highest-margin segment, gross margin expanded more than 800 basis points, and the enterprise-and-sovereign AI strategy is the correct one for HPE." (Moor Insights & Strategy).

And the backlog backs it up: more than two-thirds of HPE's AI pipeline is tied to enterprise and sovereign deployments, and HPE entered fiscal Q3 with $5.9 billion in AI systems backlog. With Juniper switching inside AMD's Heliosx rack, HPE has also gained what Moorhead calls "a quieter route into hyperscale than competing head-on." (Moor Insights & Strategy).

‍

GreenLake Intelligence: Infrastructure as an AI Operating Model

Operating AI infrastructure at enterprise scale is where many organizations struggle. HPE positioned GreenLake Intelligence as the operational layer connecting networking, compute, storage, hybrid cloud, and AI services with an agentic architecture applying identity, policy, and governance across AI agents wherever they run, plus new copilots for compute, orchestration, and observability (Futurum Group). Morpheus 9 anchors a unified CloudOps platform alongside OpsRamp and Zerto, and HPE introduced an aggressive VMware-migration offer: a free first year of Morpheus VM Essentials plus Zerto for $1 (Futurum Group). They’ve also recently partnered with ServiceNow, a move that extends GreenLake into enterprise service management (Moor Insights & Strategy).

For Alastair Cooke, these operational capabilities are becoming more important than the underlying infrastructure itself.

Alastair Cooke, Six Five Media
"Hybrid cloud success is increasingly about operational simplicity, not architectural design. HPE is showing more consistency in how it builds and operates hybrid cloud, and it supports non-HPE hardware—so customers don't need to rip and replace everything at once." (Six Five Media)

Enterprises rarely have the luxury of rebuilding their infrastructure from scratch. By supporting third-party hardware and heterogeneous environments, HPE is acknowledging the reality that most AI deployments will evolve incrementally rather than through wholesale platform replacement.

Viewed alongside the networking and AI infrastructure announcements, GreenLake Intelligence represents another piece of the same strategy. HPE isn't simply selling hardware or cloud services, it's building an operational platform designed to manage the full lifecycle of enterprise AI.

‍

Security Across Three Dimensions

As enterprise AI deployments mature, the security conversation is expanding beyond traditional cybersecurity. Protecting models and data is only part of the equation. Organizations must also secure the infrastructure AI depends on, use AI to strengthen security operations, and prepare for the new risks introduced by autonomous systems.

That broader perspective shaped HPE's security narrative at Discover 2026.

The announcements reflected a holistic approach. HPE introduced a new SASE Orchestrator designed to unify SD-WAN and Security Service Edge, along with SASE Copilot, an AI Firewall, and AI Predictive Threat Prevention running on the new quantum-safe SRX4700 platform. The company also expanded Universal ZTNA to support non-human and agentic identities, recognizing that AI agents increasingly require the same identity, access, and governance controls as human users (Futurum Group).

These announcements address different parts of the security stack and reinforce HPE's view that AI security can no longer be separated from networking, identity, or infrastructure operations. As AI agents become active participants in enterprise workflows, security policies must extend across people, machines, applications, and autonomous systems alike.

Analyst Fernando Montenegro sees this position as one of the company's emerging strengths.

Fernando Montenegro, Futurum Group
"HPE is demonstrating execution across all three dimensions of enterprise AI security—from the AI firewall on the SRX side, to SASE analyzing AI traffic flows, to Zerto's work on recovery from agentic misbehavior." (Six Five Media).

Agentic recovery is particularly noteworthy. As AI systems move from assisting users to executing workflows independently, organizations must prepare for unintended behavior from their own AI, making recovery and policy enforcement as important as prevention.

‍

The Road Ahead: From Vision to Execution

A year after closing Juniper, HPE arrived at Discover 2026 with a cohesive strategy: networking as the organizing principle, AI infrastructure as the workload layer, GreenLake as the operational control plane, and security embedded throughout (Futurum Group). Networking has become one of HPE's highest-margin businesses, AI bookings continue to grow, and much of its backlog sits in the enterprise and sovereign deployments where HPE has long-established relationships (Moor Insights & Strategy). Few vendors can deliver campus networking, data center fabrics, routing, SASE, compute, storage, hybrid cloud software, and AI infrastructure as a single portfolio — and even fewer can manage them through a common operating model while extending governance to AI agents.

David Nicholson believes that is the defining story of this year's event.

David Nicholson, Six Five Media
"HPE Discover 2026 wasn't about introducing entirely new strategies. It was about meaningful progress integrating networking, AI, hybrid cloud, and security into a cohesive enterprise platform—the kind of signal enterprises will still be talking about one or two years from now." (Six Five Media).

This observation captures what ultimately distinguished Discover 2026. HPE didn't attempt to redefine the AI market or introduce a dramatic new narrative. Instead, it demonstrated measurable progress against a strategy it has been assembling over the past several years.

For enterprise buyers, this may matter more than any single product. HPE is making the case that organizations no longer need to assemble AI platforms from disconnected vendors — they can adopt an integrated architecture backed by HPE Financial Services, partnerships with NVIDIA, AMD, and ServiceNow, and a co-sell motion with the world's largest GSIs (Futurum Group). HPE's AI story now feels less like a collection of ambitions and more like a connected platform, one built on the idea that networking isn't part of the infrastructure stack, but the foundation that makes the rest of enterprise AI possible.

‍