Dell’s new high-end Nvidia-based server is a centerpiece for its integrated AI platform aimed at enterprise customers with major AI infrastructure plans.

The Dell PowerEdge XE8812, with its core Nvidia Vera Rubin NVL4 architecture, scales up to 144 GPUs per rack and will be at the heart of the Dell AI Factory with Nvidia preconfigured package of server, storage, networking, and software infrastructure.

The Dell AI Factory with Nvidia typically includes Dell PowerEdge AI servers; Nvidia GPUs, including the H100, H200, Blackwell, and others; high-speed Ethernet or InfiniBand networking; Dell PowerScale and PowerStore storage capacity; and AI software such as Nvidia’s AI Enterprise and NIM inference microservices.

The liquid-cooled XE8812 delivers a “generational leap” in compute density and memory capacity, Dell stated. “With the shift from Nvidia GB200 NVL4 to Nvidia Vera Rubin NVL4, the platform gains expanded host memory, more cores (expanding from 144 to 176), more GPU memory, and more compute. Paired with Nvidia CUDA-X libraries this gives HPC organizations the ability to run their largest models and simulations entirely in-memory, with unparalleled processing power,” Dell stated.

The new server features 50% more memory per socket and GPU memory compared to the prior generation. The memory boost “enables organizations to run larger models and simulations entirely in-memory without the need for staging (streaming data from host memory or storage) or swapping (evicting and reloading data), both of which introduce microsecond–millisecond latency and dramatically lower effective bandwidth particularly impactful for modern AI and HPC workloads,” Dell stated.

The package includes the Integrated Dell Remote Access Controller (iDRAC), which allows IT teams to deploy, update and monitor PowerEdge servers. For rack-level visibility, the system offers the Dell Integrated Rack Controller and OpenManage Enterprise, which use real-time telemetry and automated leak detection to identify issues early, Dell stated.

“As AI and HPC simulation workloads converge, the scale and pace of these workloads are outgrowing what incremental infrastructure upgrades can keep up with,” Dell stated

The global push for AI innovation is accelerating demand for high-performance infrastructure that keeps data, compute, and control where organizations need it, Dell stated. Citing a recent Gartner study, Dell stated that as the AI growth opportunity speeds up, AI investment is projected to grow 44% year-over-year in 2026. In addition, 87% of organizations say innovation and AI are key to their business strategy.

“Building AI foundations alone will drive a 49% increase in spending on AI-optimized servers for 2026, representing 17% of total AI spending. AI infrastructure will also add $401 billion in spending in 2026 as a result of technology providers building out AI foundations,” Gartner wrote in its Forecast: AI Spending, worldwide 2024-2029 from January 2026.

Nvidia unveils Vera Rubin platform

The Dell announcement is part of a larger Nvidia rollout of its Vera Rubin architecture, which it detailed in March. Nvidia said the Vera Rubin platform combines compute, networking, and data processing into rack-scale deployments for large AI data centers. The platform integrates its Vera CPU, Rubin GPU, NVLink 6 switch, ConnectX-9 SuperNIC, BlueField-4 DPU, and Spectrum-6 Ethernet switch, along with the newly added Groq 3 LPU, into a single system designed to operate as an AI supercomputer, Nvidia stated. 

The architecture is designed to support all stages of AI workloads, from large-scale training and post-training to real-time inference. It’s aimed AI factory-type deployments or large-scale data center applications.

At the rollout, Super Micro also announced plans to roll out a Nvidia Vera Rubin-based AI server that will include up to 1,152 Nvidia Rubin GPUs and 576 Nvidia Vera CPUs in liquid-cooled racks. Its new server will be at the core of Super Micro’s Data Center Building Block Solutions (DCBBS) Blueprint offering, which defines compute, networking, advanced liquid cooling, power distribution, and site definition recommendations for building AI infrastructure, according to the vendor. 

“The DCBBS Blueprint covers the full end-to-end sequence that Supermicro has used to complete large-scale liquid-cooled projects at record-breaking speeds. On-site facility surveys conducted by the Supermicro experts assess loading dock access, data hall measurements and clearances, floor load ratings, and existing power and cooling infrastructure to inform a design proposal tailored to each project,” the vendor stated.

“The Nvidia Vera Rubin NVL4 platform is built for this convergence, and the DCBBS Blueprint for HPC defines the steps to deploy it successfully, backed by Supermicro’s proven track-record building the world’s largest liquid-cooled supercomputing clusters featuring over 100,000 GPUs,” according to Super Micro.

Nvidia