HPE adds new AI security and governance tools for enterprise deployments

HPE has unveiled new security, governance and infrastructure capabilities for its AI Factory with NVIDIA platform, aiming to help enterprises move agentic AI applications from pilot projects into production environments with greater control and operational visibility.

The announcement reflects a growing industry focus on managing AI agents that can autonomously make decisions and execute business tasks, creating new requirements for governance, security and infrastructure.

"As AI becomes more autonomous, organizations need a new architecture to run it securely, govern it responsibly and scale it economically," said Antonio Neri, president and CEO of HPE. "Across networking, servers, storage and software, HPE is delivering full-stack AI solutions with NVIDIA that build the foundation for agentic enterprises."

At the center of the update is HPE Private Cloud AI, the turnkey AI Factory platform co-developed with NVIDIA. The company is adding capabilities that allow enterprises to monitor AI agent behavior, enforce governance policies and improve the efficiency of AI inference workloads.

The platform now incorporates NVIDIA Agent Toolkit software, including Nemotron open models, NemoClaw and OpenShell secure runtime, enabling organizations to deploy AI agents with built-in policy controls and observability. HPE is also introducing secure local agent registration, allowing enterprises to approve AI models, tools and skills while maintaining centralized governance.

To strengthen resilience, HPE is extending Zerto Software to detect unauthorized or rogue AI agent actions and restore systems using continuous data protection, giving organizations a mechanism to recover from unintended AI-driven changes.

The company is also focusing on one of the biggest challenges facing enterprise AI deployments—data preparation and inference performance. HPE said built-in intelligence in its Alletra Storage MP X10000 can automatically apply metadata and governance policies to unstructured data while reducing token response times by up to 20 times. The platform is also designed to improve prompt processing efficiency and increase token throughput by up to 20%.

HPE Data Fabric Software is being enhanced with support for the Model Context Protocol (MCP) for Apache Airflow and an enterprise AI inventory that enriches distributed data with metadata, helping organizations make business data more accessible for AI workflows.

Beyond software, HPE is expanding infrastructure options for large-scale AI deployments. The AI Factory platform will support NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs, Spectrum-X Ethernet, BlueField-3 DPUs and ConnectX-8 SuperNICs, while NVIDIA Confidential Computing will add cryptographic attestation and encryption to protect AI models and enterprise data during execution.

The company said these capabilities are designed for enterprises building sovereign AI environments or deploying AI applications in regulated industries that require stronger security and compliance controls.

HPE is also introducing a new ProLiant Compute DL394 Gen12 server powered by NVIDIA Vera CPUs as an optimized platform for agentic AI and high-performance data processing, alongside unified model gateways, workload prioritization and multi-node inferencing support for up to 256 GPUs.

The announcements underscore a broader shift in the enterprise AI market, where vendors are moving beyond model development to address the operational challenges of running AI agents at scale. As organizations deploy AI into business-critical processes, governance, security, observability and infrastructure efficiency are becoming as important as model performance.

Most of the new HPE Private Cloud AI features will become available in July 2026, with additional capabilities, including agent observability, data intelligence, NVIDIA Agent Toolkit support and enhanced Zerto integration, scheduled for release in the fourth quarter of the year.