At MWC 2026, Hewlett Packard Enterprise announced a significant expansion of its service provider portfolio:

• Juniper PTX12000 Series Routers: New modular, 1.6T-ready platforms powered by Express 5 silicon. The PTX12012 supports up to 648 x 800GbE ports in a 32RU chassis, delivering up to 518.4T of total capacity.
• Juniper PTX10002 Series Routers: Fixed 2RU routers providing 14.4T or 28.8T throughput, design edfor compact AI clusters and WAN connectivity, supporting flexible 100G, 400G, and 800G port options.
• Agentic AI for Networking: Juniper Routing Director now includes an Agentic AI-ready server, allowing AI co-pilots to use natural language to analyze real-time topology, simulate “what-if” scenarios, and automate WAN configuration and troubleshooting.
• Edge Compute & vRAN: The new HPE ProLiant EL140 Gen12 server (and EL9000 chassis) handles double the traffic of previous models, built around Intel Xeon 6 SoCs with integrated AI/vRAN acceleration.
• Hardware Consolidation: Integration of the Juniper Cloud Native Router on ProLiant servers allows operators to combine RAN and routing functions on a single device.
• HPE Cloud Ops Software: A unified management stack that provides a single control plane for VMs and containers. It aims to reduce dependence on expensive hypervisors while integrating AIOps, FinOps, and security across multi-vendor environments.

Executive Summary

At MWC 2026, Hewlett Packard Enterprise announced a significant expansion of its service provider portfolio, its broadest set of announcements since closing its acquisition of Juniper Networks.

The announcements focus on three primary pillars: ultra-high-density routing for AI backbones, modular compute for the telecommunications edge, and an integrated cloud operations stack. HPE is positioning these releases to address the “bursty” and high-bandwidth nature of AI training and inference workloads that are increasingly distributed across regional data centers.

By combining Juniper’s Express 5 silicon with HPE’s ProLiant server architecture, the company provides a unified infrastructure for service providers transitioning from traditional 5G services to high-margin AI-managed services.

The Announcements

HPE’s latest releases prioritize power efficiency, port density, and the integration of “agentic AI” into WAN orchestration. The foundation of the networking updates is the Juniper Express 5 ASIC, which HPE claims provides a 49% improvement in power efficiency over previous generations.

Juniper PTX12000 Series Modular Routers

The PTX12000 family is HPE’s flagship modular routing platform for large-scale AI fabric and data center interconnect deployments. Built on the Express 5 ASIC, the platform targets super spine roles in Clos architectures and DCI backbone applications.

HPE positions these systems for environments where GPU-driven training and inference workloads require distributed, high-throughput connectivity across facilities.

Key specifications include:

• PTX12008 (8-slot chassis): Maximum system capacity of 345.6 Tbps
• PTX12012 (12-slot chassis): Maximum system capacity of 518.4 Tbps in a 32 RU form factor
• Up to 648 x 800GbE ports in the 12-slot configuration
• 43.2 Tbps line cards supporting 54 x 800GbE ports per card
• 1.6T-ready platform architecture for future line card generations
• Full 800GbE ZR/ZR+ coherent optics support across all ports (QSFP-DD and OSFP)
• Deep buffering architecture for AI traffic burst absorption
• Integrated security: line-rate MACsec encryption, DDoS protection, hardware-based integrity

Juniper PTX10002 Series Fixed-Form Routers

HPE expanded the PTX10002 fixed-form portfolio to address deployment scenarios where modular chassis are impractical. The platform targets DCI, core, data center edge, and metro aggregation roles in a compact 2 RU form factor.

Technical specifications:

• Available in 14.4 Tbps and 28.8 Tbps configurations
• Native multi-rate support: 100G, 400G, and 800G port options
• QSFP-DD and OSFP optics support including 800GbE ZR/ZR+ coherent pluggables
• 2 RU footprint for space-constrained deployments

Juniper Routing Director: Agentic AI Integration

HPE announced that Juniper Routing Director is now “agentic AI-ready,” enabling integration with external AI co-pilots via MCP. This allows Routing Director to expose WAN automation capabilities to large LLM-based systems for natural language interaction and autonomous operations.

The MCP server exposes the following data and action categories to external AI systems:

• Real-time topology and network state: live topology import, device/interface/routing observability, active monitoring and path testing data
• Service inventory and lifecycle: L2 circuits, L3VPNs, EVPN configurations; service dependency graphs; lifecycle state tracking
• Observability and compliance: configuration states, compliance audit outputs, hardware/software inventories, end-of-life tracking, MPLS path and utilization data
• Network planning and risk analysis: failure scenario simulation, what-if analysis, capacity planning data, SRLG failure impact modeling
• Historical analytics: SLA adherence metrics, outage/brownout history, MTTK and MTTR tracking

HPE claims that with appropriate permissions, AI assistants can orchestrate configuration changes, validate configurations, run active tests, optimize services, and manage security patch workflows.

HPE ProLiant Compute: Telco-Optimized Servers

HPE introduced the ProLiant Compute EL9000 chassis and EL140 Gen12 server, targeting telecommunications workloads with emphasis on 5G and AI deployments. The platform leverages Intel Xeon 6 processors with integrated acceleration features.

EL9000/EL140 Gen12 specifications:

• 72 CPU cores per server
• 24 available network ports per server
• 2U form factor with up to two servers per chassis
• Intel Xeon 6 SoC with Intel vRAN Boost and Intel Advanced Matrix Extensions (AMX)
• HPE claims the platform handles twice the network traffic compared to previous generation on a single server

Additionally, HPE announced that the Juniper Cloud Native Router is now available on both the 1U ProLiant Compute DL110 and the new 2U EL140 Gen12 servers.

Analyst’s Take

HPE’s MWC 2026 announcements are an aggressive move to position the combined HPE-Juniper portfolio as a comprehensive infrastructure provider for AI-era service provider networks. The “Networking for AI” messaging differentiates from competitors by emphasizing purpose-built infrastructure (rather than general-purpose networking adapted for AI workloads).

The announcements are also a significant milestone in HPE’s post-Juniper integration strategy, delivering a coherent portfolio that addresses the networking and compute requirements of service providers navigating the AI infrastructure transition.

The PTX12000 and PTX10002 routers position HPE competitively in high-density routing, while the agentic AI integration with Routing Director offers a differentiated approach to network automation that aligns with broader industry trends toward AI-assisted operations.

While the capital expenditure for such high-density hardware is significant and the integration of AI-driven automation carries inherent risks, the potential for service providers to consolidate hardware and reduce power consumption provides a pragmatic path toward modernization.

At the end of the day, the combination of high-density 800GbE routing, 1.6T-ready architectures, and agentic AI integration addresses real challenges in scaling AI workloads across distributed environments. HPE’s expanded portfolio provides compelling options for technology decision-makers evaluating AI infrastructure — offerings that also shows off the potential power of a combined HPE-Juniper.