Dell Technologies launched its Dell PowerEdge XE9712, a system built on Nvidia's GB200 NVL72 platform with 36 Nvidia Grace CPUs and 72 Nvidia Blackwell GPUs in a rack.
The system, available with liquid cooling, builds on Dell Technologies' AI Factory strategy for scale-out AI workloads. The Dell PowerEdge XE9712's 72 Nvidia Blackwell GPUs use NVLink to act as one GPU.
Dell said that the PowerEdge XE9712 is designed to fit into its latest rack design. The Dell Integrated Rack 7000 fits with Open Compute Project (OCP) standards and is designed to accommodate multiple generations of systems and heterogenous GPU providers. Last week, Dell launched an integrated AI system powered by AMD and its portfolio of CPUs, GPUs and software.
- Dell Technologies COO Clarke: Data gravity will drive enterprise genAI infrastructure demand
- Dell Technologies Q2 driven by AI server demand, backlog
Key facts about the Integrated Rack 7000 include:
- The 21-inch rack is designed to support CPU and GPU density.
- The rack has wider and taller server sleds for larger CPUs and GPUs.
- Dell designed the integrated rack for native liquid cooling and is capable of cooling future deployments up to 480KW.
- The rack supports Dell and third-party networking systems.
- The generative AI buildout, overcapacity and what history tells us
- AI infrastructure is the new innovation hotbed with smartphone-like release cadence
Along with the rack and new Nvidia AI system, Dell Technologies launched Dell PowerScale, Ethernet storage certified for Nvidia DGX SuperPOD.
PowerScale includes integration with RAG frameworks, Dell Data Lakehouse and an upcoming open-source document loader for Nvidia NeMo.
Dell also said it will jointly engineer systems with Intel for AI deployments with the Dell PowerEdge XE9680 and Intel Gaudi 3.
