Dell is becoming further intertwined with Nvidia with its new PowerEdge server line set to drive so-called AI factory platforms.

The XE8812, aimed at firms running intense AI and high-performance computing workloads, is a rack-scale offering capable of scaling up to 144 graphic processing units (GPUs) per rack.

Using Vera Rubin NVL4, the platform features more cores (from 144 to 176) compared to the prior generation GB200. Housed in an ORv3-style PowerRack line, Dell claims the XE8812 is “one of the industry’s densest platforms.”

“The institutions doing the world's most important research like decoding the human genome, modeling the energy systems of the future, and building the sovereign AI infrastructure that nations depend on deserve infrastructure that matches the ambition of their work,” said Arun Narayanan, Dell’s SVP for compute and networking.

“The Dell PowerEdge XE8812 reflects Dell's commitment to pushing the boundaries of what's possible, giving these organizations the density, memory and open architecture they need to tackle workloads that once seemed impossible.”

Dell’s long-running partnership with Nvidia has moved well beyond a traditional server sales relationship to one where the pair are actively co-engineering solutions for their AI factory concept, or AI-focused data centers to avoid the latter’s marketing parlance.

Their end-to-end AI infrastructure efforts have already seen the pair iterate on PowerRack, which is a rack-scale architecture integrating compute, networking, and storage into a factory-integrated system that bids to reduce deployment complexity. They also worked on Dell Exascale Storage, a unified rack management tooling designed for next-generation Nvidia platforms.

Dell is also a key proponent of Nvidia’s Spectrum-X Ethernet platforms, integrating the fabric into its broader Enterprise SONiC Distribution platform, the vendor’s open-source operating system based on the Software for Open Networking in the Cloud (SONiC).

Their latest coming together is a fanless, direct liquid cooled server line they claim is “purpose-built for the world’s most sophisticated institutions running demanding HPC and AI workloads.”

The PowerEdge XE8812s can be orchestrated using Dell’s Integrated Dell Remote Access Controller (iDRAC), while the Integrated Rack Controller and OpenManage Enterprise software tools can update cluster operators with insights on real-time telemetry and automated leak detection.

Dell goes as far as to claim that production-ready racks can be deployed and running live workloads in a little over six hours.

“The convergence of AI and HPC is redefining what organizations should expect from their infrastructure,” said Chris Marriott, VP for enterprise platforms at Nvidia. “Dell and Nvidia are raising that bar together, combining Nvidia Vera Rubin NVL4 architecture and CUDA-X libraries with Dell's engineering and at-scale deployment expertise to provide the performance, efficiency, and openness required for the world’s most demanding AI and scientific computing workloads.”