Dell Technologies World this year introduced what may be the show's most forward-looking product: Dell Deskside Agentic AI. Announced on May 18 as part of the expanded Dell AI Factory with NVIDIA, the solution allows enterprises to build, test, fine-tune, and deploy autonomous AI agents locally on Dell high-performance workstations without sending sensitive data to external cloud environments.
What Is It Solving?
The recent rise of agents (systems capable of persistent autonomous task execution) has generated enormous interest across enterprises. But the enthusiasm has been tempered by two persistent concerns: cost and data security.
Jon Seigal, SVP of Dell's client solutions group and online marketing, framed the challenge directly: enterprise IT leaders are "increasingly wary that sensitive company data could be exposed or misused in some way due to security risks, data privacy concerns, and just the lack of guardrails that are needed within OpenClaw". OpenClaw, the open-source agentic framework, allows AI agents to run autonomously on user systems, but its lack of built-in enterprise security controls has raised concerns across the cybersecurity industry.
On the cost side, AI agents capable of autonomous multi-step tasks can generate incredibly high token consumption at scale. Although per-token API prices have decreased, higher and more consistent usage makes cost forecasting challenging for internal workflows that operate all day, every day.
How Deskside Works
Dell Deskside Agentic AI pairs Dell workstation hardware with NVIDIA's NemoClaw software stack, which is specifically designed to add enterprise-grade security and governance controls on top of the open-source OpenClaw framework.
The Hardware
The solution runs on Dell's high-performance workstation lineup, including the GB10, GB 300, and Pro Precision 9 Towers. Dell says the system can handle open-weight models ranging from 30 billion to one trillion parameters locally at the desk. The Pro Precision 9 features Intel Xeon 600 processors and supports up to five NVIDIA RTX PRO Blackwell Workstation Edition GPUs.
The Software Stack: NVIDIA NemoClaw
NemoClaw is an open-source reference stack that wraps OpenClaw with enterprise controls. It consists of four components:
- OpenClaw: The agentic framework that powers persistent, autonomous, multi-step AI workflows on local hardware.
- NVIDIA Agent Toolkit: Tools for building and orchestrating long-running agents.
- NVIDIA OpenShell: A sandboxed runtime environment that allows organizations to build, deploy, and govern agents with security and privacy controls.
- NVIDIA Nemotron-3: High-performance open models for reasoning and coding.
According to Dell, guardrails can be set within NemoClaw to control what agents can and cannot do, addressing the lack of governance that has been a sticking point with vanilla OpenClaw deployments.
Dell also offers end-to-end services covering the full agentic AI lifecycle, from initial strategy and hardware deployment to workflow alignment, agent prioritization, and ongoing optimization.
Although currently pushing for agent workflows under the desk, Dell also said that this is designed to be a stepping stone for more compute.
Deskside Agentic AI is not designed to stay at the desk forever. NVIDIA OpenShell integration provides a consistent runtime from deskside workstations up to Dell PowerEdge XE servers in the data center. The NVIDIA AI-Q 2.0 Architecture adds multi-agent workflow templates for regulated industries, accelerating the path from prototype to production.
This means an enterprise can develop and validate an agent on a workstation, then deploy the same agent at scale in the data center without ever needing to re-engineer the workflow.
Enterprise Customers and Positioning
Dell highlighted several customers already deploying the AI Factory for on-premises AI workloads. Eli Lilly, Honeywell, and Samsung were cited as enterprises choosing Dell infrastructure for AI deployments where data sovereignty and cost control are critical requirements.
Dell's approach stands in contrast to cloud-native agentic AI offerings from hyperscalers, where traditionally organizations pay per-token and data flows through external infrastructure. By running agents locally, Dell argues that enterprises gain three advantages: predictable costs without metered API fees, complete data sovereignty with no IP exposure, and lower latency for workflows that interact with local systems.
Conclusion
Dell Deskside Agentic AI is a practical answer to one of enterprise AI's most pressing questions: how do you deploy autonomous agents at scale while maintaining control over costs, data, and governance? By combining Dell workstation hardware with NVIDIA's NemoClaw software stack and enterprise services, the product creates a development-to-production pipeline that keeps sensitive data on-premises throughout. For organizations that are excited about agentic AI but concerned about the risks of cloud-only approaches, Deskside Agentic AI offers a clear, available-now alternative.
References:
Dell Technologies. *Dell Technologies delivers production-ready agentic AI from deskside to data center* [Press release]. https://www.dell.com/en-us/dt/corporate/newsroom/announcements/detailpage.press-releases~usa~2026~05~dell-technologies-delivers-production-ready-agentic-ai-from-deskside-to-data-center.htm
Dell Technologies. *Dell Technologies World: A bright and beautiful road ahead* [Blog post]. https://www.dell.com/en-us/blog/dell-technologies-world-a-bright-and-beautiful-road-ahead/