New data orchestration, liquid-cooled storage, management software and validated partner solutions aim to reduce the gap between agentic AI pilots and operational deployments.
Dell Technologies has unveiled a broad set of updates to its Dell AI Factory with NVIDIA portfolio, positioning the company’s infrastructure, data platform and partner ecosystem as a foundation for enterprises moving agentic AI beyond experimentation. Announced at Supercomputing 2026 on November 16, the releases focus on making enterprise data accessible and governed for AI agents, operating large-scale AI infrastructure more efficiently, and deploying prevalidated solutions with lower integration risk.
The announcement matters because enterprise AI adoption is increasingly constrained not by model availability, but by the operational work required to connect models to trusted data, scale GPU infrastructure and maintain security, observability and governance. Dell says its approach is designed to give organizations more control by running AI workloads on infrastructure they own and manage.
Data becomes agent-ready
At the center of the update is the Dell AI Data Platform, which adds orchestration and storage capabilities intended to prepare enterprise data for AI agents. The company’s Agentic Data Gateway provides agents with a governed route to enterprise data and context, allowing teams to build integrations once rather than repeatedly connecting individual systems.
Dell says internal testing found that reusing context through the gateway reduced token consumption by up to 76% and improved answer accuracy from 60% to 94%. Those figures come from Dell’s own testing using a synthetic corpus and a specific NVIDIA model configuration, so real-world results will depend on an organization’s data, workloads and architecture.
The platform also introduces hybrid data pipeline orchestration, which can run data-processing pipelines across on-premises private clouds, Dell AI Factory environments, third-party infrastructure and public clouds. The goal is to let enterprises process large unstructured datasets where they already reside, rather than moving them into a separate AI environment.
New storage and data-intelligence features are designed to expose data provenance, quality and usage across environments containing billions of files. Dell says the platform can prepare tables, images, video and documents in a single pipeline and route vectors to an organization’s existing database, with no data migration and up to 12 times faster indexing in an Elastic benchmark using NVIDIA GPU acceleration.
Storage built for dense AI racks
Dell is expanding its Exascale storage portfolio with what it describes as the first direct-liquid-cooled storage system designed for rack-scale AI. The 2-OU platform is intended to support up to 32 NVMe drives and align with liquid-cooled AI racks, where power and cooling constraints are often central design challenges.
The company is also introducing a 1U Exascale form factor for rack-constrained environments. Dell says it offers 67% higher drive density per rack unit than the prior generation, while reducing rack footprint, power consumption and cooling requirements.
Software-defined PowerScale on Dell Exascale has also received NVIDIA-Certified Storage validation at the Foundation level. Dell positions this as a validated option for organizations seeking accelerated data access in large GPU environments.
Managing AI infrastructure at scale
Dell is introducing OpenManage Rack Scale, an infrastructure-management platform that can deploy, observe and automate physical systems from individual servers through complete racks. Dell says it can manage up to 100,000 devices, 12 times more than the previous generation, and provides a single operational view across health, alerts, firmware, inventory, power and performance.
The platform includes natural-language capabilities intended to help operations teams investigate and resolve issues without switching among multiple management tools. For enterprises building AI clusters with hundreds or thousands of accelerators, reducing operational complexity may be as important as raw compute performance.
Dell Fabric Manager adds centralized, intent-based automation and observability across multi-vendor network fabrics spanning data centers, AI environments, high-performance computing and edge sites. Dell says it automates operations from initial design and provisioning through upgrades and configuration-drift auditing, reporting up to 80% less Day-2 operational effort in internal lab analysis.
The company is also expanding its Enterprise SONiC Distribution to support additional platforms, including the Z9964F-ON and NVIDIA Spectrum SN4700 switches. The move gives customers more networking silicon choices and could reduce dependence on a single networking vendor as AI fabrics scale.
A validated route to production
Dell is pairing the infrastructure updates with an expanded Dell AI Ecosystem Program. The validated solutions cover AI security with F5 and JFrog, agentic AI with DataRobot, observability with Dynatrace, computer vision with Ipsotek, data platforms with Teradata and Iternal, knowledge assistants with Aible, Atolio and Maven AGI, coding assistance with OpenAI, and AI governance with Klique by ClearML and JFrog.
The company says these offerings have been tested on Dell AI Factory reference architectures against a consistent set of criteria, giving enterprises a more predictable route from proof of concept to production. Dell also validated Dell PowerStore and the PowerEdge XE9785 server with NVIDIA HGX B300 GPUs as part of its modular AI architecture.
Services are another part of the pitch. Dell AI Factory Engineers is an engineering-led delivery model intended to bring together platform, workload, data, governance and security expertise under one accountable team. Dell Services for AI Data Centers, meanwhile, focuses on designing facilities around the power, cooling and space requirements of dense AI deployments.
Industry context
The announcements reflect a broader shift in enterprise AI: organizations are moving from isolated model demonstrations toward systems that must access governed corporate data, operate reliably at scale and meet security and compliance requirements. Agentic AI raises the stakes because autonomous or semi-autonomous agents need reliable context, permissions and auditability—not merely access to raw data.
Dell says its AI Factory with NVIDIA portfolio now includes more than 420 offerings and is used by more than 6,500 customers. Those figures are based on Dell’s own analyses of announcements and customer-order data, but they illustrate the scale of the company’s AI infrastructure push.
“AI outcomes don’t come from models alone,” said Arthur Lewis, president of Dell’s Infrastructure Solutions Group. “They take data with the right context, infrastructure that scales and solutions validated to work together.”
NVIDIA’s Justin Boitano, vice president of enterprise software, said the companies are bringing accelerated computing and open models into AI factories so laboratories can use agents to accelerate discovery. IDC’s Ashish Nadkarni added that enterprises moving beyond experimentation increasingly need trusted data, scalable infrastructure, operational simplicity and ecosystem integration in a repeatable deployment model.
Availability
Most of the Dell AI Data Platform enhancements will begin rolling out from December 2026. OpenManage Rack Scale, Fabric Manager, several validated blueprints, the PowerEdge XE9785 with NVIDIA HGX B300 GPUs and Dell AI Factory Engineers are available now.
The Enterprise SONiC Distribution update is scheduled for January 2027, while the liquid-cooled Exascale storage platform is expected in early 2027. Dell Services for AI Data Centers and the PowerStore validation are slated for the fourth quarter of 2026.
For enterprise technology leaders, Dell’s message is straightforward: agentic AI will require more than GPUs and models. The company is betting that governed data access, high-density storage, automated infrastructure operations and prevalidated partner integrations will determine which AI initiatives reach production—and deliver measurable business value.









