On the eve of Mobile World Congress 2026, Hewlett Packard Enterprise (HPE) has decided to set the tone for telco conversations with an announcement that combines two sector obsessions: network capacity and operational efficiency. The company has unveiled new networking and compute solutions aimed at accelerating the modernization of service providers and helping them build their own Artificial Intelligence infrastructure, with low latency and high capacity “from core to edge”.
The move, announced on February 24, 2026, comes at a time when operators and cloud providers face a perfect storm: increasing traffic, virtualization becoming more costly, data sovereignty requirements, and the need to unify virtual machines and containers within the 5G core. In this context, HPE emphasizes that the challenge is no longer just deploying more hardware, but creating “AI-native” networks that can detect issues, ensure quality of service, and deliver cloud-like agility for managed services.
An integrated telco portfolio with Juniper flavor
HPE frames this announcement as another milestone in its integration with Juniper Networks, supporting a telco offering that combines routing, switching, automation, and security. The clear idea: if Artificial Intelligence will disrupt traffic patterns and raise the bar for latency, capacity, and experience, networks must respond with denser, more manageable architectures and lower energy consumption.
Rami Rahim, Executive Vice President and Chief Executive of Networking at HPE, argued that AI infrastructure has become a critical growth driver for operators, and that the company aims to help them lead “the era of AI” with networks capable of supporting complex operations, increasing data volumes, and new advanced services.
AI fabric routers: higher density, fewer redesigns, better energy efficiency
The networking announcement revolves around Juniper’s routing family, especially the PTX routers, which HPE positions as the foundation for scaling data center connectivity and building AI cluster network fabrics. A key technical highlight is the Juniper Express 5 ASIC, which claims to achieve a 49% improvement in energy efficiency over the previous generation, aiming to reduce power consumption and operating costs without sacrificing performance.
Among the features previewed for MWC 2026, HPE highlights two product lines:
- Juniper PTX12000 (modular routers): designed so operators can scale AI and cloud traffic without redesigning their infrastructure repeatedly. The company mentions 800G density in platforms prepared for 1.6T. In terms of scalability, the PTX12008 (8 slots) reaches 345.6 T, while the PTX12012 (12 slots) hits 518.4 T, enabling capacity expansion while maintaining architectural consistency and reducing latency, power, and costs.
- Juniper PTX10002 (fixed form routers): targeting high-density routing in a compact 2RU format, with 28.8 T or 14.4 T throughput and multi-speed ports 100G, 400G, and 800G. The practical message: providing room to optimize space, power, and port density as traffic grows.
Simultaneously, HPE introduces a nod to automation with Juniper Routing Director, described as “AI agent-ready”. The promise is for companies to connect it to their own AI copilots to speed up WAN routing issue resolution and simplify post-deployment operations by automating workflows and optimizing performance.
Telco compute: more traffic per server, less hardware on site
In compute, HPE presents innovations focused on speeding up 5G and AI deployments, strengthening security, and simplifying automation from edge to core. Two announcements are front and center:
- HPE ProLiant Compute EL9000 + EL140 Gen12: According to HPE, this combo enables operators to handle double the network traffic on a single server, reducing infrastructure costs and total cost of ownership. The solution features 72 CPU cores and 24 network ports per server in a 2U format, with up to two servers per chassis. It leverages Intel Xeon 6 SoC with built-in AI acceleration, including Intel vRAN Boost and Intel Advanced Matrix Extensions (AMX), as well as HPE’s security enhancements for meeting network security standards.
- Juniper Cloud Native Router on HPE servers: HPE announces availability of Juniper’s cloud-native router on the HPE ProLiant DL110 (1U) and the new HPE ProLiant EL140 Gen12 (2U). The goal is to consolidate RAN and compute functions into a single server, eliminating routing hardware at cell sites and reducing both capital expenditure and ongoing energy costs.
The sector’s takeaway is straightforward: if more functions can be packed into each rack unit — efficiently — modernization becomes less disruptive, especially in networks with thousands of sites where every device counts.
Cloud Ops Software: a “control plane” unifying virtualization, containers, and FinOps
Beyond hardware, HPE emphasizes a software layer with operational glue. HPE Cloud Ops Software is described as an integrated cloud management stack that unifies virtualization and containers, full-stack monitoring and observability, AIOps, cyber resilience, DevOps automation, and FinOps in multi-cloud and multi-vendor environments.
The message targets a current pain point: modernization pressure clashes with the high costs of certain hypervisors and the complexity of operating “telco-scale” private clouds. HPE advocates that a unified control plane helps reduce reliance on expensive technologies, lower risks, control costs, and accelerate service launches—including those driven by AI and multi-tenant managed services.
Ray Mota, CEO and principal analyst at ACG Research, supports this vision, emphasizing that AI is changing traffic patterns and increasing uplink, latency, and capacity demands. He notes that combining high-performance routing, AI-native automation, telco cloud architectures, and security is key for operators to participate “in the AI value chain” with autonomous, on-demand services.
90/9 Financing: 90 days without payments and 1% installments over 9 months
As part of the offering, HPE Financial Services launches the 90/9 Advantage program: no payments in the first 90 days, followed by ultra-competitive 1% monthly installments over the next 9 months. It’s presented as a mechanism to accelerate AI projects and modernization efforts amid uncertain commodity pricing, applicable to networking, compute, storage, and software portfolios.
MWC 2026: demos and narrative—automation to cut costs and generate AI revenue
HPE anticipates showcasing live demos at MWC 2026 that demonstrate how to automate operations, cut costs, boost AI-driven revenue, and speed up deployment times. The company mentions its presence at stand 3N10 and at the HPE meeting center located at the VIP entrance of Hall 3, emphasizing that the message extends beyond products: it’s about positioning. With operators redefining their roles, HPE aims to convey that “AI-native” infrastructure isn’t an add-on but the pathway to compete effectively.
FAQs
What advantage does an 800G network offer for AI clusters in data centers?
It enables increased capacity and throughput with fewer links and less complexity, facilitating denser network fabrics for AI workloads and cloud traffic, while promising predictable latency as volume grows.
What does it mean for a router or platform to be “agentic-AI ready” in WAN operations?
It means the management software can connect to AI copilots or agents to automate diagnostics and fixes, speeding up incident resolution and reducing manual work on repetitive operational tasks.
Why are operators seeking to unify virtual machines and containers in the 5G core?
Because managing two separate environments costs more and complicates operations. Unification simplifies deployments, observability, automation, and cost control in services that need rapid scaling.
How does consolidating RAN and routing functions onto a single server reduce costs at cell sites?
It can eliminate dedicated routing hardware, lower energy consumption, and simplify maintenance, especially in remote locations where each additional device adds CAPEX, OPEX, and operational complexity.
via: hpe