• Home
• Technology

 Hewlett Packard Enterprise (HPE) and NVIDIA have announced major enhancements to the HPE AI Factory with NVIDIA portfolio.

Unveiled at NVIDIA GTC Berlin 2026, these updates aim to enable enterprises to securely deploy and manage Agentic AI applications across hybrid cloud environments and sovereign infrastructures, while enhancing data security, performance, and operational control.

Sherri Williams, Senior Vice President and General Manager of Private and Flexible Cloud Solutions at HPE, highlighted that scaling AI agents requires an infrastructure that ensures data sovereignty and control. She noted that HPE is providing the governance and operational tools needed to move applications from trial phases into full production.

Expanding HPE AI Essentials and Streamlining Developer Experience

HPE is enhancing HPE Private Cloud AI capabilities by updating the HPE AI Essentials platform. This provides developers, independent software vendors (ISVs), and AI agents with secure access to model, inference, and knowledge services, seamlessly integrating them into applications via APIs and Model Context Protocol (MCP) servers.

Key security and operational features include:

Access Control: Role- and project-based access management.

Auditability & Logging: Comprehensive tracking of usage logs and support for machine/system identities.

Automated Documentation: OpenAPI standard integration to automatically generate technical documentation, software development kits (SDKs), and code samples to accelerate development.

Unified Computing Resources and Edge AI Deployment

HPE Private Cloud AI now supports running NVIDIA H200 and NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs within a single unified compute cluster. This comes alongside the general availability of NVIDIA’s open-source KAI Scheduler, which optimizes GPU resource allocation through quotas and priorities to enhance workload management.

In the industrial sector, French energy management and automation leader Schneider Electric plans to leverage the enhanced HPE Private Cloud AI solution to support industrial AI agents at the edge. This enables real-time analytics and decision-making closer to operational sites while retaining control over sensitive data.

High-Density Computing and Reference Architectures for System Protection

To further empower AI factory infrastructure, HPE unveiled the liquid-cooled NVIDIA Vera CPU Rack by HPE, featuring 256 NVIDIA Vera CPUs and supporting up to 22,500 isolated execution environments. The system offers up to 80% faster performance in CPU-based agentic AI tasks.

HPE also announced support for the NVIDIA OpenShell Reference Architecture for Safe AI Agents, designed to build, monitor, and secure AI agents. This architecture combines:

NVIDIA OpenShell secure runtime environment on NVIDIA Vera processors.

NVIDIA Sentry technology with the NVIDIA BlueField-4 DPU for hardware-enforced security controls.

NVIDIA Vera Rubin NVL72 system for advanced inference processing.

Additionally, HPE will support the NVIDIA DSX POD reference design, utilizing pre-fabricated modular infrastructure units to accelerate the deployment of large-scale AI factories.

Confidential Computing and Availability Timeline

To protect sensitive data and models, HPE is expanding support for NVIDIA Confidential Computing on HPE ProLiant Compute DL380a Gen12 servers powered by NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs. This delivers hardware-backed trusted execution environments to isolate AI workloads across virtual machines, containers, and bare-metal deployments.

Late 2026: Rollout of HPE AI Essentials updates allowing administrators to grant developers API access.

2027: Launch of a self-service developer portal for registering applications, discovering APIs and MCP services, and managing credentials.