At HPE Discover 2026, HPE announced a series of enhancements to its AI infrastructure portfolio to help enterprises operationalize agentic AI with greater governance, security, and scalability. The updates expand HPE AI Factory offerings within the NVIDIA AI Computing by HPE portfolio and focus on moving AI initiatives from pilot projects into production environments.
The announcements center on HPE Private Cloud AI and larger-scale HPE AI Factory deployments, adding new capabilities for agent governance, data preparation, inference efficiency, and confidential computing.
“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, helping customers move from experimentation to production with control and confidence.”
“Every layer of the computing stack is being reinvented for the age of AI agents,” said Jensen Huang, founder and CEO of NVIDIA. “Together with HPE, we are building AI factories for this new era of computing, powered by NVIDIA Vera CPUs, accelerated infrastructure, and secure AI software, to help enterprises transform their data into intelligent action.”
HPE Private Cloud AI Adds Agentic AI Capabilities
HPE Private Cloud AI, the company’s turnkey AI platform co-engineered with NVIDIA, is receiving several enhancements focused on enterprise deployment of AI agents.
A key addition is support for the NVIDIA Agent Toolkit, which includes NVIDIA Nemotron open models, NVIDIA NemoClaw, and the NVIDIA OpenShell runtime environment. Together, these technologies provide a framework for agent reasoning, policy enforcement, behavioral monitoring, and operational governance.
HPE is also introducing the HPE ProLiant Compute DL394 Gen12 with NVIDIA Vera CPU support as a compute platform optimized for agentic AI workloads and high-performance data processing.
To improve operational resilience, HPE is extending HPE Zerto capabilities to monitor agent actions and identify potentially harmful or unauthorized behavior. Continuous data protection features will allow organizations to restore environments to known-good states when necessary.
The platform also supports local agent registration, enabling enterprises to approve AI models, tools, and skills through centralized governance and security policies.
Addressing Data Preparation and Inference Efficiency
Data readiness remains a significant challenge for enterprise AI deployments, and HPE is targeting this area with several updates.
Built-in intelligence in the HPE Alletra Storage MP X10000 enables automatic metadata tagging and enforcement of governance policies for unstructured data. HPE says the approach helps organizations prepare AI-ready datasets more quickly while significantly improving inference performance.
The company reports that token response times can be reduced by up to 20x, while prompt-processing efficiency and overall token throughput can improve by up to 20%.
HPE Data Fabric Software is also expanding support for agentic AI workflows. New capabilities include Model Context Protocol (MCP) support for Apache Airflow and an enterprise AI inventory that enriches distributed datasets with metadata, improving discoverability and governance.
For organizations seeking simpler deployment models, HPE will also offer a standalone HPE Data Fabric appliance running on HPE ProLiant servers.
Scaling AI Infrastructure and Managing Costs
HPE is introducing several capabilities designed to improve resource utilization and control operational costs in large AI environments.
A new unified model gateway provides governed access to multiple AI models through a centralized interface. Additional features include workload prioritization and multi-node inference support that can scale across up to 256 GPUs.
The platform also supports fine-tuning pre-trained models, including NVIDIA Nemotron models, using enterprise data through NVIDIA NeMo integration.
These enhancements are intended to help organizations maximize GPU utilization, manage token consumption costs, and support the long-term growth of AI infrastructure.
Confidential Computing Comes to HPE AI Factory
For large-scale and sovereign AI deployments, HPE is integrating NVIDIA Confidential Computing technologies across its AI Factory portfolio.
The technology protects AI models and sensitive data during runtime through hardware-backed security, encryption, and cryptographic attestation. HPE said the capability is designed to help organizations meet regulatory, industry, and sovereignty requirements while maintaining operational performance.
Additional security capabilities leverage NVIDIA BlueField DPUs and NVIDIA DOCA software to provide zero-trust enforcement, runtime threat detection, and encrypted networking across AI infrastructure environments.
The enhancements will be available across HPE AI Factory at Scale and HPE Sovereign AI Factory deployments.
Expanded NVIDIA Hardware Integration
HPE also announced broader support for NVIDIA’s latest AI infrastructure technologies across its AI Factory portfolio.
HPE AI Factory solutions now support NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs, NVIDIA Spectrum-X Ethernet networking, NVIDIA BlueField-3 DPUs, and NVIDIA ConnectX-8 SuperNICs.
Based on NVIDIA reference architectures, the solutions are designed to support a range of AI use cases from model development and training through production-scale deployment. The platforms also integrate NVIDIA AI Enterprise software and ecosystem offerings from HPE’s Unleash AI partner program.
Availability
HPE said the new unified model gateway and additional HPE Private Cloud AI capabilities will be available in July 2026.
HPE Data Fabric Software updates are scheduled for October 2026.
Additional Private Cloud AI capabilities, including agentic observability, data intelligence services, HPE Alletra Storage MP X10000 integration, NVIDIA Agent Toolkit support, and NVIDIA NemoClaw support, are expected in the fourth quarter of 2026.
HPE Zerto support for agent monitoring and recovery workflows is planned for the fourth quarter of 2026, alongside the availability of NVIDIA Confidential Computing across HPE AI Factory solutions.
Support for NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs, Spectrum-X Ethernet, BlueField-3 DPUs, and ConnectX-8 SuperNICs is available immediately.
Amazon