Hewlett Packard Enterprise teams with Nvidia to enhance secure AI deployment

Hewlett Packard Enterprise and Nvidia announced a significant expansion of their ongoing partnership on June 16, 2026, at HPE Discover Las Vegas. The collaboration targets deploying AI agents at scale without losing control of data, compliance posture, or hybrid and fully on-premises infrastructure.

The centerpiece of the announcement is a set of new capabilities inside HPE Private Cloud AI, built specifically for regulated industries and enterprises running hybrid or fully on-premises infrastructure. New features are expected to begin rolling out in July 2026, with additional capabilities targeted for Q4 2026.

What is actually being built

HPE is integrating Nvidia’s Agent Toolkit into Private Cloud AI, giving organizations a framework for building and orchestrating agentic AI workflows inside their own controlled environments.

On the hardware side, the partnership introduces the HPE ProLiant Compute DL394 Gen12 server, powered by Nvidia’s Vera CPUs, purpose-built compute that does not require enterprises to route sensitive data through public cloud infrastructure.

The security architecture draws on Nvidia Confidential Computing, which creates protected execution environments for AI models and data. For industries like healthcare, financial services, and government contracting, that level of isolation is not a nice-to-have.

HPE is also adding Zerto, its continuous data protection product, specifically to address rogue agent actions. Agentic AI systems that can take autonomous steps, write to databases, and trigger workflows can cause real damage if they go sideways. Zerto provides a rollback mechanism for what the AI has done to your environment.

Why this moment matters

The HPE and Nvidia relationship goes back further than this announcement. The two companies launched NVIDIA AI Computing by HPE in 2024, then expanded the collaboration in 2025. What is different now is the maturity of the offering. Earlier iterations were largely about getting GPU-accelerated infrastructure into enterprise data centers. This iteration is about making the AI that runs on that infrastructure trustworthy enough to use in production, in regulated settings.

Data sovereignty regulations in the European Union, financial services rules in the US, and classified computing requirements in government mean that a meaningful slice of enterprise AI demand simply cannot live in a hyperscaler’s cloud. HPE has historically served that customer base. Adding Nvidia’s AI stack to that proven distribution channel gives both companies a credible answer for segments that AWS, Azure, and Google Cloud cannot easily serve.

Competitive stakes and what to watch

HPE and Nvidia are not operating in a vacuum here. Dell Technologies and Super Micro Computer are both active in the enterprise AI server market and have their own Nvidia partnerships. The differentiation HPE is reaching for is not just the hardware, it is the integrated stack: compute, security, data protection, and orchestration bundled into a turnkey solution rather than a collection of components a customer has to assemble themselves.