Dell expands NVIDIA AI Factory to boost enterprise ROI
Dell Technologies has announced updates to its Dell AI Factory with NVIDIA, expanding the portfolio across data platforms, infrastructure, and services as companies move beyond AI pilots and face tougher scrutiny over return on investment.
More than 4,000 customers have deployed the Dell AI Factory with NVIDIA, according to Dell. The company also cited analysis showing early adopters achieving up to 2.6x ROI within the first year.
The announcements reflect a wider shift in enterprise AI. Many CIOs now build AI systems in-house and deploy them on-premises as tools reduce the cost and time needed to develop applications. This approach increases the focus on owned infrastructure and on data preparation for AI workloads.
Dell framed unclear ROI as the main barrier to scaling AI deployments. It outlined three requirements it considers central to measurable returns: data platforms that make enterprise information AI-ready, infrastructure that scales from pilot to production, and services that reduce operational complexity during rollout.
Data platform
The update includes changes to the Dell AI Data Platform with NVIDIA, which Dell describes as the data foundation of the AI Factory offering. The platform combines Dell storage with modular data engines and NVIDIA computing, networking, and software.
It is designed to support a range of AI data and inference patterns, including retrieval-augmented generation and multimodal search. It also targets agentic workflows and large-scale data processing.
Dell said the latest updates are intended to make it faster and easier for organisations to turn internal data into AI results. Pricing was not disclosed.
Desktop to data centre
Dell also outlined new infrastructure options for AI development on desktops and workstations, as well as production deployments in the data centre. It positioned the line-up as support for workflows from prototyping through to scaled inference and training.
On the desktop side, Dell introduced Dell Pro Max with GB10 and Dell Pro Max with GB300. It described them as desktop AI supercomputers designed to build and run AI, and said it is the first OEM to ship a desktop system with the NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip.
Dell said this desktop configuration delivers up to 20 petaFLOPS of FP4 performance and 748GB of coherent memory.
For workstations, Dell pointed to Dell Pro Precision systems aimed at developers and data scientists. It said tower models can be configured with up to five NVIDIA RTX PRO Blackwell Desktop Generation GPUs, and highlighted mobile workstation configurations using NVIDIA RTX PRO Blackwell Generation Laptop GPUs.
In the data centre, Dell named PowerEdge XE9812 as its flagship liquid-cooled server and said it uses the NVIDIA Vera Rubin NVL72 platform for real-time training and inference. It also announced PowerEdge XE9880L, XE9882L, and XE9885L, described as liquid-cooled servers featuring NVIDIA HGX Rubin NVL8.
For general-purpose servers that need AI acceleration, Dell said PowerEdge R770, R7715, and R7725 can be configured with the NVIDIA RTX PRO 4500 Blackwell Server Edition GPU. It also introduced PowerEdge R9822 and M9822 with NVIDIA Vera CPU options.
Networking and racks
The infrastructure update also included networking products and rack-level systems. Dell said the PowerSwitch SN6000-series are NVIDIA Spectrum-6 Ethernet switches, and highlighted 1.6Tbs throughput as well as liquid cooling and co-packaged optics options.
Dell also said PowerSwitch SN5610 and SN2201 now offer additional network operating system choices, including Cumulus Linux and Enterprise SONiC Distribution by Dell Technologies. It also referenced NVIDIA Quantum-X800 InfiniBand Q3300-LD liquid-cooled switches for high-bandwidth networking.
Integrated Rack Scalable Systems now include Dell PowerSwitch and NVIDIA liquid-cooled switching, according to the company. Dell positioned this as a way to provide unified rack-level power and cooling management for AI infrastructure.
Dell also highlighted support for NVIDIA NVQLink and CUDA-Q. It said it is the first OEM to integrate NVIDIA NVQLink with CUDA-Q across PowerEdge servers featuring NVIDIA AI infrastructure, targeting quantum-classical computing research use cases.
Services and blueprints
On the software and services side, Dell said it has updated Dell AI Solutions with a modular architecture. It also referenced Dell Automation Platform blueprints and NVIDIA AI Enterprise software as part of its packaged approach.
Dell listed several offerings aimed at specific workloads, including a knowledge assistant and a ClearML blueprint focused on GPU cluster management and workload scheduling. It also described an agentic AI platform created with partners including Cohere's North, DataRobot, and NVIDIA, which it said is designed for secure deployment and management of AI agents with orchestration, governance, and observability.
Dell also announced Dell Accelerator Services for Agentic AI, describing it as packaged support spanning experimentation, validation, and integration.
Michael Dell, Chairman and Chief Executive Officer of Dell Technologies, said:
"Two years ago, enterprises were asking how to access AI technology. Today, they're asking how to make their data AI-ready, how to operationalise AI at scale and how to prove ROI. The Dell AI Factory with NVIDIA answers all three questions. We're brought in from the start as a trusted advisor, helping customers navigate their entire AI journey-from turning raw data into AI fuel, through deployment and to measurable business outcomes."
Jensen Huang, Founder and Chief Executive Officer of NVIDIA, said:
"AI infrastructure is being built everywhere - every company will be powered by it, every country will build it- and it demands integrated data platforms, scalable infrastructure and deployment expertise. Dell Technologies delivers all three, with NVIDIA at the core. The Dell AI Factory with NVIDIA is a proven infrastructure blueprint for every phase of AI powering the next industrial era."