Dell expands AI Factory with NVIDIA for agentic AI workloads

By Sophie Lindqvist

What has Dell expanded?

More than 6,500 customers have adopted Dell AI Factory with NVIDIA, Dell said as it expanded the platform with software, deskside systems and data-centre infrastructure for agentic AI.

Dell positioned the update around a shift from pilot projects to broader deployment across development teams, office environments and core infrastructure. The company said enterprises want more control over data, operations and costs as they deploy AI agents more widely.

Dell said the expanded package combines validated models and software, managed services, desktop and workstation configurations, modular data-centre designs and new Dell PowerEdge servers. The company designed that package for organizations running AI workloads on infrastructure they own.

Dell's portfolio spans several levels rather than one appliance. The announcement places developer desktops at one end and racks, managed services and communications infrastructure at the other.

Takeaway: Dell is presenting a production path that connects local developer systems with managed infrastructure through a coordinated hardware and software package.

What changed in the software stack?

Dell said it validated Reflection's Beam open-weight model on Dell AI Factory with NVIDIA. The company also added or updated ecosystem programme solutions in three areas:

• AI observability: Dynatrace
• Knowledge assistant and agentic AI platform: Mistral
• Computer vision: Fogsphere

Dell said SUSE Linux will be added as an operating system option for the stack. The company also outlined a layered security approach that combines Dell infrastructure with NVIDIA OpenShell Reference Architecture for runtime governance and Red Hat AI Factory software for orchestration.

Dell introduced Dell Managed Services for AI to cover the operational work around those deployments. According to Dell, the service helps customers govern access to AI resources, manage consumption and monitor costs as projects move from early trials into broader operation.

Takeaway: The software additions focus on observability, governance, orchestration and cost control, not just model access.

How does the deskside range work?

Dell expanded its deskside agentic AI offering because it expects AI workloads to split between central infrastructure and systems used by individual developers and teams. The company said new validated reference deployment templates for Dell Deskside Agentic AI Solutions for NVIDIA specify the hardware, models and software stack.

The templates cover:

• Agentic coding
• Agentic VSS
• NVIDIA NemoClaw
• Hermes Agent on NemoClaw
• A higher education AI lab platform

Dell said the templates can scale from one developer system to deployments of 1,000 or more systems. Both ends of that range use the same reference-deployment approach.

For individual developers and smaller teams, Dell introduced a 64GB unified memory configuration for Dell Pro Precision with GB10. Dell said the desktop can be linked in clusters of two or four systems to support larger models and more complex workflows.

Dell also said Dell Pro Precision with GB300 will be offered with Windows. The company said the configuration brings up to 748GB of coherent memory and 20 petaflops of FP4 compute to deskside environments already managed by enterprise IT teams.

Takeaway: The deskside range now covers one developer system through deployments of 1,000 or more, with local hardware tied to enterprise management.

What changes in the data centre?

At the data centre, Dell announced a modular reference design called Dell Modular Data Centre with NVIDIA DSX POD. Dell said the containerized design includes compute, storage, networking, rack-level power and liquid cooling.

Dell said the design can connect to prepared power infrastructure. The company said this approach reduces on-site integration work.

The Dell PowerEdge M9882 is a CPU-focused system that places 256 NVIDIA Vera CPUs in a single 48U rack. Dell said the system supports orchestration, tool use, code execution and evaluation alongside GPU-based inference systems.

Dell said the aim is to separate CPU resources from GPU clusters rather than rely on mixed-use servers. The company said mixed-use servers can complicate power and cooling management.

The Dell PowerEdge R8822 targets conventional 19-inch rack environments. Dell said it supports agentic AI tasks, reinforcement learning, data processing and analytics, as well as up to four double-wide GPUs.

Dell announced the Dell PowerEdge XE9810 for rack-scale inference. According to Dell, the server uses 16 direct-liquid-cooled compute trays powered by Groq 3 Language Processing Units for high-throughput inference in production AI deployments.

Dell also said it is working with NVIDIA on Aerial RAN Computer platforms running on Dell PowerEdge servers. The company said the systems target communications service providers that want radio access network and AI workloads on a shared compute footprint rather than separate systems.

Dell's three server platforms divide the rack portfolio by workload. The range covers CPU-heavy orchestration and tool use, conventional accelerator expansion and dedicated high-throughput inference.

Takeaway: Dell's rack portfolio groups CPU-heavy orchestration, flexible GPU systems and dedicated inference into distinct products.

Why are enterprise buyers central to this launch?

Michael Dell, chairman and chief executive officer of Dell, tied the release to changing infrastructure budgets. He said:

"Enterprises used to think about AI infrastructure as an IT budget question, but that's changed. AI infrastructure is no longer a cost center. It's a value creation center."

"Companies are expanding and reallocating budgets because AI infrastructure drives growth, productivity and competitive advantage. Together with NVIDIA, we are giving enterprises the integrated platforms they need to move agentic AI from experimentation to results, on infrastructure they own and control."

Jensen Huang, founder and chief executive officer of NVIDIA, framed the hardware requirement in more direct terms. He said:

"AI agents reason, use tools and learn-and they need a new kind of computer."

"Together, Dell and NVIDIA are building the computing foundation for the age of AI-from the deskside to the data center-so companies can build and run AI agents securely at industrial scale."

For enterprise buyers, the decision now covers more than accelerator hardware. Dell's package also addresses runtime governance, operating-system choice, service access, consumption and cost monitoring, putting procurement and operations alongside compute.

Matt Kimball, vice president and principal analyst at Moor Insights & Strategy, focused on the market's shift toward measurable output. He said:

"The market is entering a new phase where customers are focused less on building AI infrastructure and more on generating measurable business outcomes."

"The continued expansion of the Dell AI Factory with NVIDIA reflects a mature approach that combines validated solutions, ecosystem integration, automation and services, helping organizations move from pilots to production with greater confidence and less risk."

Kimball's assessment puts pilot-to-production conversion beside the hardware rollout. Dell's additions of validated stacks and managed services address that same operational transition, according to the company.

Takeaway: The sales pitch now rests on business outcomes, infrastructure control and a defined path from trials to production.

What should buyers watch next?

Turning more of the 6,500-plus customer base into repeatable production deployments is the next proof point for Dell and NVIDIA. Dell says the rollout now combines infrastructure ownership, runtime governance and cost monitoring, the controls it believes enterprises need as agent use widens.

Takeaway: The next test is repeatable production across the 6,500-plus customer base, with data, operations and costs still under enterprise control.