“NVIDIA DSX POD — a reference design for complete NVIDIA Vera Rubin AI factories built by partners inside modular containers, integrated and validated before delivery — will reduce deployment time at a prepared site from six months to as little as 14 days, with a global ecosystem of partners delivering DSX POD systems worldwide starting in early 2027.”
Each pod houses seven liquid-cooled Vera Rubin NVL72 racks (504 Rubin GPUs) in a road-transportable ~28,000 kg container; named build partners include Bull, Dell, HPE, Lenovo, Foxconn (Ingrasys), Gigabyte, Quanta, Pegatron, Schneider Electric and Vertiv. Resolves on whether partner deliveries begin in the first part of 2027.