Hewlett Packard Enterprise has expanded its AI-native networking portfolio as part of its broader plan to bring self-driving network operations to hybrid environments. The announcement, made on 4 December at HPE Discover Barcelona 2025, marks one of the company’s most significant developments since completing its Juniper Networks acquisition five months earlier.
HPE has unified new AIOps features, common hardware, and updated software across HPE Aruba Networking and HPE Juniper Networking. The aim is to deliver a consistent operational experience and improve visibility across compute, storage, networking, and cloud systems. These enhancements are built on HPE’s agentic AI framework, which supports autonomous operations and integrates more tightly with GreenLake Intelligence.
Rami Rahim, executive vice president, president and general manager of Networking at HPE, said customers now require networks engineered for AI-driven environments. He said, “In the era of AI, customers need networks that are purpose-built with AI and for AI to handle the rapid growth of connected devices, complex environments, and increasing security threats.” He added that HPE intends to redefine user experiences by offering autonomous, high-performing networks across varied environments.
HPE has blended capabilities from Aruba Networking Central and Juniper’s Mist platform. The Mist Large Experience Model, which analyses data from applications such as Zoom and Teams alongside digital twin simulations, will become available in Aruba Networking Central. Aruba’s Agentic Mesh technology is being added to Mist to strengthen anomaly detection and root-cause analysis. Mist will also adopt organisational insights and NOC features from Aruba’s platform. To support deployment flexibility, new WiFi-7 access points will work across both management systems.
The latest version of Aruba Networking Central On-Premises introduces a redesigned interface and builds on earlier AIOps capabilities. It brings intelligent remediation, actionable AI alerts, client insights, proactive monitoring, and simplified documentation search into one on-premises environment.
New data centre and edge networking hardware
HPE has broadened its hardware line-up with products designed for performance-intensive AI workloads, especially as inferencing moves to the edge. The company introduced the HPE Juniper Networking QFX5250 switch, the first OEM switch built on Broadcom’s Tomahawk 6 silicon. It offers 102.4Tbps of bandwidth and is positioned as an Ultra Ethernet Transport-ready switch for connecting GPUs in AI data centres. It draws on HPE’s liquid cooling technology, Junos innovations, and AIOps features to offer performance, energy efficiency, and ease of operation.
To complement its data centre hardware, HPE introduced the Juniper Networking MX301 multiservice edge router. The compact 1RU device delivers 1.6Tbps performance and 400G connectivity. It is designed to support AI inferencing closer to where data is created and is suited for multiservice edge routing, enterprise deployments, metro networks, and mobile backhaul.
HPE outlined new solutions developed with NVIDIA and AMD that extend the company’s AI factory networking capabilities. These include high-speed, secure, low-latency connectivity for linking users, devices, and AI agents to AI factories, as well as long-haul data centre interconnect between clusters and clouds. HPE highlighted its continued work with NVIDIA’s Spectrum-X Ethernet networking platform and BlueField-3 DPUs to support production AI environments.
HPE also detailed its collaboration with AMD on the “Helios” AI rack-scale architecture. The solution supports trillion-parameter AI training workloads, offers 260TB/s of scale-up bandwidth, and delivers 2.9 exaflops of FP4 performance. It includes a purpose-built HPE Juniper scale-up switch developed with Broadcom, designed to accelerate AI training and inference using standards-based Ethernet.
Strengthening hybrid cloud AIOps and simplifying IT operations
HPE is advancing its hybrid cloud AIOps strategy through enhancements to HPE OpsRamp Software. The updates bring together telemetry from HPE Compute Ops Management, Aruba Networking Central, and Juniper Networking Apstra, giving IT teams a unified view of their environments.
New features include deeper GreenLake integration, full-stack observability through Apstra Data Center Director and Data Center Assurance, and predictive assurance tools for faster issue identification and resolution. HPE has also introduced capabilities such as Compute Copilot, self-service root-cause analysis, and OpsRamp integration for simplified troubleshooting and improved operator experiences.
The company is rolling out agentic root-causing tools, Model Context Protocol support for no-code integrations, and new AI agents in GreenLake Intelligence. Together, these enhancements aim to reduce blind spots, bridge data silos, and provide guided actions across the IT landscape.
To support adoption of these AI-native networking technologies, HPE Financial Services has announced two new financing programmes. The first offers zero-percent financing for networking AIOps software purchased through term-based licences. The second provides the equivalent of 10 percent cash savings for leasing networking hardware designed for AI workloads. An optional multi-OEM take-out service is available for customers replacing legacy equipment.
HPE confirmed availability timelines for the newly announced products. The QFX5250 switch will be available in Q1 2026, while the MX301 router will be released in December 2025. OpsRamp integrations and related updates will be rolled out between late 2025 and mid-2026.
