NVIDIA MGX
A modular server architecture specification introduced by NVIDIA that standardizes chassis, power delivery, liquid cooling and motherboard layout for GPU servers. It supports high-density AI computing deployments of tens to hundreds of kilowatts per rack and accelerates the design and deployment of AI servers.
Timeline (last 90 days)
DG Matrix released the Interport SST Cell and Interport OS, providing a modular power conversion and software control layer for its Interport 360 Flex system aimed at the NVIDIA MGX architecture.
NVIDIA MGX is a modular server architecture specification introduced by NVIDIA, intended to make GPU servers from different vendors follow a unified standard in chassis, power supply, liquid cooling, and motherboard layout.
MGX racks typically integrate multiple high-power GPUs, with a single rack's power reaching tens or even hundreds of kilowatts, far exceeding the density of traditional data center racks.
All Facts (3)
NVIDIA MGX is a modular server architecture specification introduced by NVIDIA, intended to make GPU servers from different vendors follow a unified standard in chassis, power supply, liquid cooling, and motherboard layout.
50%UnverifiedMGX racks typically integrate multiple high-power GPUs, with a single rack's power reaching tens or even hundreds of kilowatts, far exceeding the density of traditional data center racks.
50%UnverifiedDG Matrix released the Interport SST Cell and Interport OS, providing a modular power conversion and software control layer for its Interport 360 Flex system aimed at the NVIDIA MGX architecture.
50%