Particle.news

HPE Unveils Unified Self-Driving Network Stack for AI Data Centers and Edge

HPE says the new stack will speed AI deployments by reducing GPU idle time through automated operations that tighten security.

Overview

  • HPE announced on June 16, 2026 that it is rolling out an expanded self-driving networking portfolio that unifies management across campus, edge, data center, and AI factory environments.
  • The company integrated its HPE Networking CX switches into the HPE Mist platform and extended Marvis agentic AIOps to data centers and AI factories to detect problems, diagnose root cause, and take trusted remediation actions automatically.
  • HPE introduced two HPE Juniper Networking QFX switches — the QFX5140 for inference clusters and the QFX5252 scale-up tray for AMD Helios racks — which are designed to cut network latency and keep GPUs busy during training and inference.
  • A new AI-native SASE offering combines SD-WAN and Security Service Edge (SSE) in one console to simplify policy management, speed zero-trust rollouts, and reduce attack surface for distributed AI workloads.
  • HPE Financial Services launched a Network Migration Program with hardware and 0% software financing plus an asset-reuse option to lower cost and risk for customers moving to the AI-ready stack, a push that builds on HPE’s 2025 acquisition of Juniper and aims to move enterprises from AI pilots into production.