Dell expands AI Factory with storage, compute and governance tools

Dell Technologies is expanding its Dell AI Factory with NVIDIA with persistent storage, additional compute options and supported software aimed at moving agentic AI from pilots into production. The additions are designed to help enterprise teams manage application state, scale workloads and operate AI systems under greater control.
- Dell added PowerStore persistent storage to Dell AI Factory with NVIDIA.
- The PowerEdge XE9785 adds an AMD-based platform with eight NVIDIA HGX B300 GPUs.
- NVIDIA OpenShell provides runtime governance through policy-enforced sandboxes.
- Dell Supported NVIDIA AI Enterprise includes Dell support for covered software components.
- PowerStore, the XE9785 and the listed software additions target production agentic AI.
Persistent storage for production agents
Dell says new Dell PowerStore storage provides the Dell AI Factory with NVIDIA modular foundation with persistent, low-latency storage for AI-enabled databases, metadata, vector stores and Kubernetes services.
That storage is intended to keep application state available as services restart, update or scale. In production, AI applications can depend on databases, indexes, queues, transaction logs and recovery information, rather than only on access to source data. Dell says maintaining this state can help agents keep track of work, respond consistently and recover from disruption.
PowerStore works alongside PowerScale and ObjectScale as complementary storage tiers. It can be used with Dell AI Factory with NVIDIA today, while PowerStore guidance is expected to be added to the modular foundation design guide later this year.
More compute choice for demanding workloads
Dell is also adding the Dell PowerEdge XE9785 to the modular architecture. The system uses an AMD-based CPU, an air-cooled platform and eight NVIDIA HGX B300 GPUs.
The XE9785 complements the PowerEdge XE9780, giving customers another platform choice within a common foundation. Dell says organizations can use the added flexibility to align compute with workloads such as reasoning, training, fine-tuning, long-context agentic AI and agentic inference.
The Dell PowerEdge XE9785 is available now for the Dell AI Factory with NVIDIA modular foundation. Dell describes the architecture as a validated, ready-to-deploy full-stack solution that brings computing infrastructure, storage, networking and AI software together. Predefined modules can be added as demand grows, including expanded compute, networking or storage without re-architecting the entire foundation.
Layered controls for AI operations
As agents use tools and take actions, Dell says operational security must extend beyond prompts and model-level safeguards. NVIDIA OpenShell, part of the NVIDIA Open Agent Safety Platform alongside the NVIDIA Sentry reference system design, adds runtime governance through policy-enforced sandboxes.
These sandboxes can control actions, tool access, filesystem activity and network egress. OpenShell also provides ways to trace agent behavior and define boundaries for interactions with enterprise systems. Red Hat AI Factory software adds platform capabilities for building, deploying, managing and securing AI across hybrid environments, including isolation, network policy, role-based access and secrets management through Red Hat OpenShift.
Dell Supported NVIDIA AI Enterprise combines NVIDIA AI Enterprise software, including NVIDIA NIM, NeMo microservices and NVIDIA Blueprints, with Dell support. Dell provides Level 1 and Level 2 support for covered software components. Dell says a single support path can reduce coordination and troubleshooting friction as organizations scale AI capabilities. Dell Supported NVIDIA AI Enterprise, the NVIDIA OpenShell reference design and Red Hat AI Factory software are available now.
For users, the additions target the operational problems that emerge after an AI demonstration: preserving application state, matching infrastructure to workload growth and governing how agents interact with systems. Dell's approach is to combine those capabilities in a modular foundation intended to help enterprises move toward production while retaining control of their data, models and AI operations.
Dell Technologies: Move Enterprise AI From Pilots to Production With Confidence


