Jensen Huang sends in the robot ducks to sell Nvidia's AI-everywhere vision
At Nvidia's GTC keynote in Berlin, Jensen Huang announced DSX POD, a modular reference design for factory-built AI infrastructure. The system integrates seven liquid-cooled Vera Rubin compute racks with Spectrum-X networking in standardized containers. Partners will integrate and test units before delivery, with availability expected in early 2027. Nvidia claims deployment timelines could shrink from six months to 14 days at prepared sites. The system uses closed-loop liquid cooling with coolant temperatures up to 45°C and outdoor dry coolers, with Huang claiming potential water savings of 54 million liters daily at 2-gigawatt scale.
Key points
- DSX POD combines seven Vera Rubin compute racks with Spectrum-X networking
- Modular containerized design aims to reduce deployment time to 14 days
- Water-efficient liquid cooling design claims 54 million liters daily savings at 2GW
Summary and key points written by AI from the headline and a short excerpt, so they can be incomplete or wrong. The Register did the reporting; read the original for the full story.
Read the full story at The Register
