HPE expands AI infrastructure for agentic workloads

Hewlett Packard Enterprise (HPE) is positioning itself to capitalize on the growing demand for agentic AI infrastructure, announcing a series of updates to its AI datacenter portfolio ahead of its Discover event in Las Vegas. The enhancements integrate technology from its recent Juniper Networks acquisition and deepen collaboration with Nvidia, targeting enterprises preparing for AI-driven automation and workflows.

The company’s revamped offerings include Private Cloud AI, AI Factory at-scale, and Sovereign AI Factory, each tailored to different deployment models. Private Cloud AI, a turnkey solution co-developed with Nvidia, now incorporates Nvidia’s Nemotron models, NemoClaw security stack, and OpenShell secure runtime to create an agent operating environment. It also features HPE’s Alletra Storage MP X10000 device, previously previewed at the company’s December 2025 Discover event. These updates will be available from July 2026.

For larger-scale deployments, HPE’s AI Factory at-scale and Sovereign AI Factory now support Nvidia Confidential Computing, designed to protect user models and private data during workload execution. This feature is slated for release in the fourth quarter of 2026. The platforms also integrate Nvidia’s RTX PRO 6000 Blackwell Server Edition GPUs, Spectrum-X networking, BlueField-3 DPUs, and ConnectX-NICs, aiming to optimize performance for AI workloads.

Networking and software integrations

HPE’s integration of Juniper networking technology, acquired in late 2025, plays a central role in the updated portfolio. The company has cross-pollinated capabilities from its HPE MIST and Aruba Central AIOps platforms, enabling Marvis AI actions in Aruba Central and CX switch management via Mist. This unification is intended to simplify operations and provide a consistent experience across compute, storage, and networking layers.

New hardware announcements include the Juniper QFX5140, a 16 Tbps switch built on the Trident5 chipset, targeting inference clusters and edge AI deployments. The QFX5252 Switch tray for AMD Helios, a scale-up module for AMD’s rack-scale AI platform, was also introduced. Additionally, HPE has consolidated its SD-WAN and Security Service Edge (SSE) offerings into a single console, Networking EdgeConnect, to streamline management and policy enforcement.

Broader strategy and market context

HPE’s updates reflect a broader industry shift toward agentic AI, where autonomous systems interact with enterprise data and workflows. The company’s emphasis on security, scalability, and cross-domain management aligns with customer demands for infrastructure that can support both cloud and on-premises AI workloads. By leveraging Nvidia’s latest GPUs and its own Data Fabric Software, HPE aims to enable agents to access and process data across hybrid environments while maintaining security and compliance.

The integration of Juniper’s networking portfolio into HPE’s GreenLake platform further underscores the company’s push for a unified operating experience. This move is likely to appeal to enterprises seeking to simplify the deployment and management of AI and traditional workloads across their infrastructure stacks.

What to watch

The success of HPE’s strategy will depend on how quickly enterprises adopt agentic AI and whether the company can differentiate its offerings from competitors like Dell, Lenovo, and hyperscalers. The rollout of Confidential Computing in Q4 2026 and the performance of new hardware like the QFX5140 switch will be key milestones to monitor. Additionally, HPE’s ability to maintain momentum in cross-platform integration, particularly between its networking and AI operations, will be critical as it competes in an increasingly crowded market.