LAS VEGAS – Hewlett Packard Enterprise (HPE) used this week’s Discover 2026 event to continue pouring AI-powered updates and features into the market. The latest focuses on Nvidia-based efforts around AI Factory and Private Cloud AI Platform, a move aimed at accelerating enterprise returns on AI-powered investments.
HPE’s AI Factory and Sovereign AI Factory platforms are based on Nvidia’s reference architecture and include Nvidia’s AI Enterprise software and HPE Unleash AI program partners. These platforms will have immediate access to Nvidia’s RTX Pro Blackwell Server Edition graphics processing units (GPUs), Spectrum-X Ethernet, BlueField-3 digital processing units (DPUs), and ConnectX-8 SuperNICs (network interface cards).
The platform also integrates Nvidia’s Confidential Computing service, available through HPE Services. The Nvidia service, first launched in mid-2023 and rolled out in early 2024, is designed to protect AI models and private data while running on-premises or sovereign deployments. It uses cryptographic attestation and encryption to establish a “chain of trust” that protects hardware, software, and datasets.
HPE’s Private Cloud AI also runs Nvidia architecture and gains support for secure local agent registration. This allows customers to approve AI models, skills, and tools that meet centralized governance and security policies. This is increasingly important for sovereignty and edge-based AI use cases. You’ll also get insights from HPE’s Alletra Storage MP X1000. This allows customers to automatically apply metadata and governance policies to prepare data for AI applications and increase token throughput. This is a concept that comes amid growing concerns about token usage.
There’s also a new unified model gateway for managed frontier model access, active workload prioritization, and multimode inference for up to 256 graphics GPUs. It can support fine-tuning pre-trained AI models, such as Nvidia Nemotron open models that help control token costs.
The private cloud AI product will also gain support for Nvidia’s Agent Toolkit and NemoClaw later this year, and will be available on the vendor’s ProLiant Compute DL394 Gen12 hardware next year. This combination runs Nvidia’s Vera central processing units (CPUs) as the “compute optimization foundation” for agent AI and data processing.
These Nvidia-focused updates come on the heels of HPE tying the chip giant’s Vera Rubin products to its rack-scale, compute, and Cray supercomputing platforms, including the last platform to gain Quantum-X800 InfiniBand support for the Cray GX5000.
HPE recently announced that HPE Rack-Scale Systems with NVIDIA Vera Rubin NVL72, HPE Compute We announced that we are expanding our factory solutions. GX5000.
Data Fabric, Zerto Gain Feature
Other new updates include HPE’s Data Fabric software, which extends support for Model Context Protocol (MCP) to Apache Airflow, and access to Enterprise AI Inventory, which can inject metadata into distributed data sets to power agent workflows. There are also standalone Data Fabric appliances that run on HPE’s ProLiant Compute servers to help simplify deployment.
HPE’s Data Protection Zerto software gains support for agent action monitoring and continuous data protection. This allows users to identify when an agent goes rogue and uses continuous data protection that allows data to be rewound when an anomaly occurs.
These updates follow a number of new features that HPE has been rolling out over the past few months, starting with a major effort at the Europe-based Discover event late last year and continuing to deliver new features and services into the year.
