Details
- NVIDIA introduced DSX POD, a reference design for complete Vera Rubin AI factories integrated and validated inside modular containers before delivery.
- NVIDIA is working with Anduril, Armada, Bull, Crusoe, Dell Technologies, HPE, Lenovo, Schneider Electric, Vertiv and other partners to manufacture and deploy the systems worldwide in early 2027.
- Each pod integrates computing, networking, cabling, rack-level power and liquid cooling, arriving ready to connect to a prepared site’s power, cooling and network infrastructure.
- Factory-based integration shifts specialized labor, construction planning and functional validation away from the deployment site, potentially reducing installation time from six months to as little as 14 days.
- DSX POD extends NVIDIA’s broader DSX platform, which combines reference designs, software, infrastructure technologies and partner systems for designing and operating AI factories.
Impact
By separating factory integration from site preparation, DSX POD could accelerate the expansion of AI cloud and sovereign-compute capacity while reducing exposure to onsite delays and integration errors. Over the next 12–24 months, the model may steer infrastructure investment toward standardized, repeatable designs and strengthen demand for partners supplying power, cooling, networking and modular data-center systems.