Dell targets enterprise AI execution gap with local agentic AI systems and integrated AI infrastructure

Dell Technologies Inc. today is kicking off its Dell Technologies World conference by expanding its artificial intelligence portfolio with enhancements aimed at helping enterprises move AI projects from experimentation into large-scale production, with a particular focus on agentic AI, data orchestration, rack-scale infrastructure and on-premises deployment.

The announcements build on the company’s Dell AI Factory with Nvidia initiative, which Dell said now has more than 5,000 customers globally.

The centerpiece of the news is Dell Deskside Agentic AI, a new offering that combines Dell workstations, Nvidia Corp.’s NemoClaw software stack, and Dell services to enable enterprises to develop and run AI agents locally rather than relying exclusively on cloud infrastructure.

Dell executives framed the announcements as a response to mounting enterprise frustration over the complexity and cost of deploying generative and agentic AI systems at scale.

“Most enterprises don’t have an AI ambition problem,” said Sam Grocott, senior vice president of product marketing at Dell Technologies, during a media briefing. “They have an AI execution problem.”

‘Execution problem’

He said enterprises are increasingly struggling with data management, energy consumption, sovereignty concerns and escalating cloud costs as AI workloads expand. In response, customers are demanding infrastructure that spans “from the desk side to the data center, to the edge, and all the way out to the cloud.”

The new deskside offering is designed to address growing concerns about the economics of agentic AI. Unlike conventional chatbots, agentic systems can autonomously execute multi-step workflows and continuously consume inference tokens, which are the chunks of data large language models use to generate results. That can create potentially enormous cloud bills.

“We had a single developer burn through 1 billion tokens in 24 hours,” said Jon Siegal, senior vice president of Dell’s client solutions group. “That was a $3,400 cloud bill.”

Dell said the deskside systems allow organizations to run open-weight AI models locally while keeping sensitive data inside their own environments. That translates into spending reductions of as much as 87% compared with using public cloud over a two-year period.

The offering includes several hardware configurations, ranging from compact Dell Pro Max systems for smaller models to high-end workstation towers capable of supporting models with up to 1 trillion parameters. Dell is also integrating Nvidia OpenShell across the Dell AI Factory portfolio to provide what Siegel called “a secure sandbox for running, building, testing and fine-tuning agents” locally.

Addressing fragmented data

Beyond agentic AI, Dell also announced a series of updates to its Dell AI Data Platform (pictured, stack view) intended to help enterprises prepare and manage data for AI applications.

Many organizations remain stuck in pilot phases because their data is fragmented across silos and difficult to govern, said Varun Chhabra, senior vice president of infrastructure and telecom marketing.

“Most customers are not short on AI ideas,” he said. “One of the hidden bottlenecks that we find when we talk to customers is the ability to get their data strategy right.”

Among the updates are enhanced orchestration and search capabilities that Dell said can index billions of unstructured files and accelerate vector indexing up to 12-fold. The company also announced graphics processing unit-accelerated SQL analytics capabilities developed with Nvidia and Starburst Data Inc. that promise up to six times faster query performance on Nvidia Blackwell graphics processors.

Dell also unveiled support for Nvidia Omniverse, a scalable development platform for physical AI, digital twins, and 3D simulation. The integration enables customers to combine enterprise storage and semantic search capabilities with digital twin and physical AI workflows.

Turnkey AI

On the infrastructure side, Dell introduced PowerRack, a turnkey rack-scale system that integrates computing, networking, storage, cooling and management into pre-engineered units for AI and high-performance computing deployments.

Chhabra said customers increasingly want integrated systems instead of assembling components from multiple vendors.

“With Dell PowerRack, customers no longer need to buy components and hope they work together,” he said.

The company also announced new cooling systems, including PowerCool CDU C7000, which it described as the first rack-mount cooling distribution unit capable of supporting Nvidia’s upcoming Vera Rubin NVL72 platform in a compact 4U form factor.

Dell is also using the event to further expand its ecosystem partnerships as well. New collaborations include integrations with Google LLC, OpenAI LLC, Palantir Technologies Inc., ServiceNow Inc. and Hugging Face Inc..

Under the new partnerships, Dell said Google Gemini models will be available through Google Distributed Cloud on Dell infrastructure, while OpenAI’s Codex coding agent will integrate with Dell’s AI Data Platform and AI Factory infrastructure. Palantir’s Foundry and Artificial Intelligence Platforms are also being brought on-premises to Dell infrastructure.

A new Dell AI Ecosystem program will validate partner applications on Dell infrastructure and accelerate deployment. “It gives them that reach and presence in the market by working through Dell,” said Caitlin Gordon, vice president of product management for private cloud and AI solutions at Dell

Most of the new announced products and integrations will become available throughout 2026, with some offerings shipping immediately.

Image: Dell