Dell unveils PowerEdge XE8812 AI/HPC server with up to 144 GPUs per rack

Dell Technologies introduced the PowerEdge XE8812 server at a June 24, 2026 event in Madrid. The rack‑scale system uses NVIDIA’s Vera Rubin NVL4 architecture and can house up to 144 GPUs per rack, delivering more than 300 kW of power. It is liquid‑cooled, fan‑less, and built on the open‑rack OCP standard, offering roughly 50 % more memory per socket and GPU memory than the previous generation. Integrated management tools such as iDRAC, Dell Integrated Rack Controller and OpenManage Enterprise provide real‑time telemetry and automated leak detection.

Early deployments include the US Department of Energy’s Doudna supercomputer at Lawrence Berkeley National Laboratory, France’s InstaDeep Kyber cluster, the UK Wellcome Sanger Institute’s genome‑decoding infrastructure, and Australia’s Monash University MAVERIC platform for cancer, climate and genomics research. Dell plans a worldwide rollout of the server in early 2027, positioning it as a turnkey solution for demanding AI and high‑performance computing workloads.