Dell Technologies Advances Enterprise Agentic AI from the Deskside to the Data Centre with NVIDIA
Dell Technologies has announced a broad range of advancements across the Dell AI Factory with NVIDIA, covering AI solutions, ecosystem innovations, deskside agentic AI and purpose-built infrastructure. The developments aim to help enterprises move agentic AI initiatives from experimentation to production using infrastructure they own and control.
More than 6,500 customers have now adopted the Dell AI Factory with NVIDIA, up from 5,000 in the previous quarter, highlighting growing enterprise demand for an integrated approach to AI deployment.
Why it matters
Enterprise agentic AI is expected to become the largest AI workload by 2028, with its growth projected to outpace generative AI by nearly three times by 2030. However, increased adoption alone does not guarantee business results. The economics of AI, including the cost of inference and the growing consumption of tokens, will play a critical role in determining its commercial value.
Inference costs are projected to decline by more than 90% by 2030, while agentic AI is expected to drive a 24-fold increase in token consumption. As organisations look to maximise their investments, uncertainty over return on investment (ROI) remains the leading obstacle for 59% of enterprise AI deployments. Meanwhile, 83% of organisations identify end-to-end vendor support across the AI lifecycle as their top priority when selecting an AI partner.
Owning AI infrastructure gives enterprises greater control over costs, data and operations than public-cloud-only approaches. According to Dell, the Dell AI Factory can be up to 90% more cost-effective than public cloud services over two years.
However, infrastructure ownership and cost savings alone are insufficient. Organisations also need a foundation that enables them to select appropriate tools, monitor performance, enforce policies and manage expenditure as deployments expand.
The Dell AI Factory with NVIDIA combines an extensive AI portfolio with validated solutions, an expanding open ecosystem, expert services and an end-to-end AI-ready data foundation. The approach is designed to help enterprises progress from pilot projects to production with greater confidence.
Validated AI solutions and services accelerate outcomes
Delivering business value from agentic AI requires more than a capable model. The Dell AI Factory with NVIDIA brings together Dell-engineered systems, software and automation with a validated ecosystem of enterprise AI software. The integrated approach is intended to help organisations adopt proven AI solutions more quickly and with less risk, while retaining control of their infrastructure.
Reflection’s Beam open-weight model has been validated on the Dell AI Factory with NVIDIA. The Apache 2.0-licensed model is designed for software engineering and long-horizon agentic tasks, enabling organisations to deploy it on infrastructure they control.
Additional validated solutions and updates are being introduced through the Dell AI Ecosystem Program for enterprise applications, including AI observability with Dynatrace, a knowledge assistant and agentic AI platform with Mistral, and computer vision with Fogsphere.
With the addition of SUSE Linux, the Dell AI Factory with NVIDIA also expands the range of choices available across the AI software stack, from models and applications to operating systems. Security is another essential consideration as enterprises deploy autonomous agents at scale. Dell Technologies is developing a layered security architecture built on the Dell AI Factory with NVIDIA that incorporates Red Hat AI Factory software.
Configurations can combine validated, hardware-rooted infrastructure with the NVIDIA OpenShell Reference Architecture for runtime governance and Red Hat AI Factory software for platform orchestration. The integrated approach spans runtime, network, data and infrastructure layers, helping organisations operate autonomous AI agents securely across enterprise environments.
Beyond deployment, Dell Managed Services for AI is designed to help organisations optimise resources, govern access, and manage AI consumption and costs as they move from pilot projects to production. Development teams gain secure self-service access to AI resources, while IT departments retain the governance and financial controls required to manage deployments.
Agentic AI extends to the deskside
As AI becomes more integrated into everyday work, deskside computing is emerging as an important component of enterprise AI architecture. Dell notes that 81% of organisations are already planning, piloting or deploying AI-enabled PCs.
To address this demand, Dell is expanding Dell Deskside Agentic AI with new validated reference deployment templates. Dell Deskside Agentic AI Solutions for NVIDIA provide blueprints designed to simplify the deployment and scaling of agentic AI workflows.
Each blueprint specifies the hardware, AI models and software stack required for use cases such as agentic coding, agentic VSS, NVIDIA NemoClaw, Hermes Agent on NemoClaw and a higher-education AI lab platform. The templates are intended to help organisations scale securely from individual developers to deployments spanning 1,000 or more systems.
Dell is also expanding its deskside hardware portfolio. Dell Pro Precision with GB10 will be available in a 64GB unified memory configuration aimed at individual developers and smaller teams. The system provides an entry point for local agentic AI development, with the capacity to run leading open-weight models.
Two or four systems can be connected in clusters to support larger models, additional agents and more complex AI workflows. Dell Pro Precision with GB300 will also be available with Windows, bringing up to 748GB of coherent memory and 20 petaflops of FP4 computing performance to the deskside. The system is designed to operate within the Windows environments enterprises already manage.
Purpose-built infrastructure for AI at scale
As agentic AI and other AI workloads expand, Dell is introducing new platforms designed to deliver the dense, efficient computing capacity required for large-scale deployments. These additions aim to provide organisations with greater flexibility while addressing the growing infrastructure demands associated with AI token consumption.
Modular data centres with integrated AI infrastructure
As demand for AI capacity increases, the ability to deploy infrastructure quickly is becoming a competitive consideration. Building on its experience in modular data centre design, Dell is introducing the Dell Modular Data Center with NVIDIA DSX POD reference design.
The containerised solution is designed to provide a complete AI foundation, incorporating compute, storage, networking, rack-level power and liquid cooling. Organisations can connect the system to prepared power infrastructure, reducing the complexity of on-site integration.
Dell said the solution draws on more than 20 years of its engineering expertise in integrating, testing, delivering, deploying and managing infrastructure throughout its lifecycle.
CPU-based execution for agentic AI
The Dell PowerEdge M9882 server integrates 256 NVIDIA Vera CPUs into a single 48U rack. It is designed for agentic workloads involving orchestration, tool use, code execution and evaluation, operating alongside GPU infrastructure used for model inference.
Separating CPU compute from GPU clusters allows organisations to deploy the NVIDIA Vera CPU Rack to make better use of available data centre power and cooling capacity, rather than over-provisioning mixed-use servers. This approach is intended to simplify infrastructure management as deployments scale.
Agentic AI for mainstream enterprise environments
The Dell PowerEdge R8822 server brings air-cooled, two-socket NVIDIA Vera CPU performance to standard 19-inch rack environments. It is optimised for agentic AI workloads such as agent orchestration and sandboxing, as well as reinforcement learning, data processing and analytics.
The platform supports up to four double-wide GPUs to accelerate AI and other compute-intensive workloads.
High-throughput inference at scale
The Dell PowerEdge XE9810 server is designed to deliver rack-scale inference through 16 direct-liquid-cooled compute trays powered by Groq 3 Language Processing Units (LPUs).
NVIDIA LPU-based inference is positioned as a complementary, disaggregated option alongside existing GPU clusters, addressing the high-throughput requirements of production agentic AI workloads.
Bringing AI to the network edge
Telecommunications networks are increasingly becoming AI-enabled environments in their own right. As mobile AI traffic grows, the NVIDIA Aerial RAN Computer (ARC) platforms running on Dell PowerEdge servers will provide communications service providers with a blueprint for processing AI workloads at the network edge.
Michael Dell, chairman and chief executive officer, Dell Technologies, 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.”
The architecture allows radio access network (RAN) and AI workloads to share the same compute infrastructure instead of operating on separate, isolated systems. This is designed to lower total cost of ownership and simplify operations.
Jensen Huang, founder and chief executive officer, NVIDIA, 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.”
Reflection’s Beam model is now available on the Dell AI Factory with NVIDIA. AI observability with Dynatrace will be available on the Dell AI Factory with NVIDIA starting in Q4 CY26. The knowledge assistant and agentic AI platform with Mistral are now available on the Dell AI Factory with NVIDIA.
Computer vision with Fogsphere will be available on the Dell AI Factory with NVIDIA starting in Q4 CY26. SUSE Linux will be available as an operating system for the Dell AI Factory with NVIDIA starting in early Q1 CY27. Dell Managed Services for AI will be available in Q1 CY27.
Dell Deskside Agentic AI Solutions for NVIDIA are now available on GitHub. Dell Pro Precision with GB10 in the 64GB configuration will be available from 23 October Dell Pro Precision with GB300 running Windows is coming soon.
The Dell Modular Data Center with NVIDIA DSX Reference Design will be available in early 2027. The Dell PowerEdge M9882 will be available in Q1 CY27. The Dell PowerEdge R8822 will be available this month. The Dell PowerEdge XE9810 will be available in Q1 CY27. The NVIDIA Aerial RAN Computer (ARC) framework on Dell PowerEdge servers is coming soon.
