HPE Expands AI Factory Portfolio for Agentic AI Deployments

Enterprises are moving beyond AI experimentation toward production environments that require strict governance and security. At HPE Discover 2026, HPE announced enhancements to its AI infrastructure portfolio designed to operationalize agentic AI with greater scalability and control.

The expansion focuses on the HPE AI Factory offerings within the NVIDIA AI Computing by HPE portfolio. These updates target the transition of AI initiatives from pilot projects into functional, large-scale deployments. The new capabilities address critical bottlenecks in agent governance, data preparation, inference efficiency, and confidential computing.

This shift signals a move toward autonomous systems that require a fundamental reinvention of the computing stack. As organizations deploy more autonomous agents, the architecture must support secure execution and economic scaling across networking, servers, storage, and software.

A central component of this strategy is the HPE Private Cloud AI platform, co-engineered with NVIDIA. The platform is receiving several enhancements to facilitate the enterprise deployment of AI agents. This includes support for the NVIDIA Agent Toolkit, which integrates NVIDIA Nemotron open models, NVIDIA NemoClaw, and the NVIDIA OpenShell runtime environment. These technologies establish a framework for reasoning, policy enforcement, behavioral monitoring, and operational governance.

To handle the compute demands of these workloads, HPE is introducing the HPE ProLiant Compute DL394 Gen12. This platform includes support for NVIDIA Vera CPUs and is optimized for agentic AI workloads and high-performance data processing.

Operational resilience is also a primary focus. HPE is extending HPE Zerto capabilities to monitor agent actions and identify unauthorized or potentially harmful behavior. Continuous data protection features will allow organizations to restore environments to known-good states when necessary. Furthermore, the platform supports local agent registration, allowing enterprises to approve AI models, tools, and skills through centralized governance.

The integration of these technologies suggests that the next phase of enterprise AI will be defined by how well companies can govern autonomous actions rather than just how many models they can train.

How will your organization manage the security risks of an increasingly autonomous agentic workforce?