Nvidia has officially released its DGX Station workstation PC that the company unveiled last year during GTC 2025. The new system is targeted at software developers, researchers, data scientists, and anyone who needs more AI horsepower than what Nvidia's smaller DGX Spark is capable of. Nvidia states that DGX Station systems are now available to order and will begin shipping in the coming months from Asus, Dell, Gigabyte, MSI, Supermicro, and HP.
DGX Station takes advantage of Nvidia's latest GB300 Grace Blackwell Ultra Desktop Superchip that combines a 72-core Grace CPU with a Blackwell Ultra GPU linked together with a 900 GB/s NVLink C2C interface. The system comes armed with a whopping 748GB of total onboard memory; the CPU portion is paired to 496GB of LPDDR5X rated at 396GB/s of bandwidth, and the GPU is paired with 252GB of HBM3e memory rated at 7.1 TB/s of bandwidth. Both memory pools are unified, allowing the CPU and GPU to share each other's memory for maximum AI performance.
Just like a typical workstation PC, Nvidia has armed the DGX Station with three PCIe Gen 5 x16 slots, one wired with 16 lanes and eight lanes for the other two. The workstation system officially supports discrete GPU options to plug into its PCIe slots for extra tasks such as simulation and ray-traced visualization.
Supported GPUs are the RTX Pro 6000 Workstation Edition, RTX Pro 6000 Blackwell Max-Q Workstation Edition, RTX Pro 4000 Blackwell SFF Edition, and RTX Pro 2000 Blackwell graphics cards. There are also four M.2 slots, audio connectors, and USB ports.
The workstation uses Nvidia's ConnectX-8 SuperNIC for networking, which supports speeds of up to 800 Gb/s through two QSFP112 ports. The system is designed to accelerate AI projects by hooking up to two DGX Stations together to scale model capacity and performance. Power connectivity comes in the form of a single 24-pin ATX power connector, a single 8-pin EPS connector, and three 12V-2x6 power connectors for the GPU, to feed the system's 1600W official power rating.
