Dell Deskside Agentic AI Cuts Cloud Costs by 87%
Dell Technologies has introduced Dell Deskside Agentic AI at Dell Technologies World in Las Vegas. The numbers are hard to ignore. The new solution lets workgroups deploy and scale agentic AI workflows locally. Importantly, it removes the cost, latency, and data sovereignty constraints tied to cloud-only approaches.
The announcement directly responds to a growing enterprise pain point. As AI workflows shift toward agentic architectures, token usage compounds fast. As a result, cloud costs can quickly spiral out of control, even as token prices fall. One Dell developer, for example, ran up a $3,400 bill by burning through one billion tokens in just 24 hours.
Dell Deskside Agentic AI directly tackles this problem. Organizations can break even against public cloud API costs in as little as three months. Moreover, over two years, they can cut spending by up to 87% compared to cloud APIs.
The solution runs across a range of Dell high-performance workstations. Each system is sized for different workloads and budgets. The Dell Pro Max with GB10 is compact and power-efficient. It handles prototyping with models from 30 billion to 200 billion parameters. The Dell Pro Precision 9, on the other hand, is an enterprise tower. It features Intel Xeon 600 processors and up to five NVIDIA RTX PRO Blackwell GPUs. It supports models from 30 billion to 500 billion parameters. At the top end, the Dell Pro Max with GB300 handles frontier-level models, from 120 billion up to 1 trillion parameters. It runs on the NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip.
Crucially, Dell Deskside Agentic AI does not work in isolation. With NVIDIA OpenShell now supported across the Dell AI Factory with NVIDIA, enterprises get a single security and policy layer. That layer spans from deskside workstations all the way to Dell PowerEdge XE servers. In addition, NVIDIA AI-Q 2.0 blueprint support gives organizations a tested foundation. Teams can, therefore, deploy multi-agent workflows for research, decision support, and complex tasks. Consequently, this accelerates the move from pilot to production.
Furthermore, Dell is promoting the Dell-NVIDIA AI-Q 2.0 Reference Architecture. It is powered by the Dell AI Data Platform. Specifically, it targets regulated industries, financial services, manufacturing, and the public sector. These are environments that need tighter control over data and operations.
Dell COO Jeff Clarke put it plainly. “The most efficient token is the one produced closest to the data,” he said. “Most enterprise data isn’t in the cloud. Dell Deskside Agentic AI gives every workgroup a secure local environment. It keeps costs predictable and IP inside the building. What works at the desk scales to the data center. That’s a deployment model for the next decade.”
Similarly, NVIDIA’s Justin Boitano, Vice President of AI Platforms, noted that enterprises can develop locally and scale securely. With NVIDIA OpenShell, they can also deploy Dell Deskside Agentic AI on one consistent platform.
Meanwhile, industry analyst Ryan Shrout, President of Signal65, said the solution bridges the gap between local control and enterprise scalability. It also provides room for continued iteration, all while maintaining data sovereignty and predictable costs.
Dell has already rolled out the AI Factory to more than 5,000 customers worldwide. As a result, with Dell Deskside Agentic AI now available, the company is clearly betting enterprises are ready to move beyond cloud dependency.
Dell Deskside Agentic AI is available now. NVIDIA OpenShell, on Canonical Ubuntu and Red Hat AI, and NVIDIA AI-Q 2.0 for Dell AI Factory with NVIDIA are also available now.
