HPE expands self-driving networking for AI factories

HPE has expanded its self-driving networking strategy across AI factories, data centres and the enterprise edge, targeting increasingly distributed, AI-driven environments.

The announcement accompanies a broader HPE update on new AI infrastructure and software developed with NVIDIA to help customers move AI agents into production with tighter governance, security and operational control.

As organisations adopt autonomous AI systems and automated workflows, they are under pressure to rethink network and infrastructure design. HPE is positioning self-driving networking as part of a wider full-stack approach spanning networking, servers, storage and software.

Andrew Fox, general manager of HPE Networking Australia and New Zealand, said customers in the region want networks that require less manual intervention as AI workloads grow.

"Enterprises can no longer rely on traditional networking models to support the demands of agentic AI and autonomous workflows," Fox said.

"Customers across Australia and New Zealand are looking for self-driving networks that can anticipate issues, adapt in real time and resolve problems independently as AI workloads scale. By combining AI-native operations, unified security and full-stack infrastructure, we are enabling organisations to move faster from experimentation to production with greater control, performance and trust."

AI push

The broader product update centres on HPE Private Cloud AI, which HPE described as a turnkey AI factory system co-engineered with NVIDIA. New functions are intended to help businesses deploy AI agents with stronger oversight, improve data preparation and manage the cost of running large models.

Additions include support for NVIDIA Agent Toolkit software, including NVIDIA Nemotron open models, NVIDIA NemoClaw and the NVIDIA OpenShell secure runtime. These tools are designed to provide a framework for monitoring agent behaviour, applying policy controls and reducing deployment risk.

HPE is also adding HPE ProLiant Compute DL394 Gen12 with NVIDIA Vera CPU to HPE Private Cloud AI as a compute-focused platform for agentic AI and data processing. HPE Zerto Software will also gain features designed to identify rogue agent actions and let customers revert to an earlier clean state through continuous data protection.

Data management is another focus of the update. HPE said HPE Private Cloud AI can convert unstructured data into AI-ready pipelines more quickly, while HPE Alletra Storage MP X10000 can apply metadata and governance policies to prepare information for AI applications.

According to HPE, the storage system can cut token response times by up to 20 times, while HPE Private Cloud AI can improve prompt processing efficiency and increase token throughput by up to 20%. HPE Data Fabric Software is also being extended to support model context protocol workflows in Apache Airflow and to add an enterprise AI inventory for distributed data environments.

Security focus

Beyond the private cloud offering, HPE is strengthening security in larger-scale and sovereign AI factory deployments. Changes include integrating NVIDIA Confidential Computing through HPE Services for on-premises and sovereign installations.

HPE said the confidential computing setup is designed to protect models and private data during execution using cryptographic attestation and encryption across hardware, software and datasets. This structure is intended to help customers meet regional and industry requirements.

The broader HPE AI Factory will also include NVIDIA BlueField and NVIDIA DOCA for zero-trust policy enforcement, runtime threat detection and networking encryption. These measures are intended to protect AI workloads, agents and data across the environment.

At the hardware level, HPE said its AI Factory at scale and Sovereign AI Factory will be offered with NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs, NVIDIA Spectrum-X Ethernet, NVIDIA BlueField-3 DPUs and NVIDIA ConnectX-8 SuperNICs. The platform is based on NVIDIA reference architectures and is intended to support use cases ranging from development to production deployments.

Antonio Neri, president and chief executive officer of HPE, linked the update to changing requirements as AI systems become less dependent on direct human control.

"As AI becomes more autonomous, organisations need a new architecture to run it securely, govern it responsibly, and scale it economically," Neri said.

"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."

NVIDIA framed the partnership as part of a broader shift in enterprise computing.

"Every layer of the computing stack is being reinvented for the age of AI agents," said Jensen Huang, founder and chief executive officer 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 said HPE AI Factory with NVIDIA RTX PRO Blackwell Server Edition GPUs, Spectrum-X Ethernet, BlueField-3 DPUs and ConnectX-8 SuperNICs is available now.