Huaying Technology Logo
NVIDIA DGX GB300
AI Computing / Data Center

NVIDIA DGX GB300

The NVIDIA DGX GB300 is a rack-scale liquid-cooled AI system built on the NVIDIA Grace Blackwell Ultra superchip. It integrates 72 Blackwell Ultra GPUs and 36 Grace CPUs, delivering 2,592 Arm Neoverse V2 cores, up to 20TB of GPU memory (with bandwidth up to 576TB/s), and 37TB of total fast memory. With FP4 Tensor Core performance reaching 1440 PFLOPS (sparse), it offers up to 1.5x denser FP4 performance and 2x faster attention mechanisms compared to the DGX GB200. Designed with the NVIDIA MGX chassis, the DGX GB300 can scale to thousands of nodes when combined with DGX SuperPOD. Preloaded with NVIDIA Mission Control and AI Enterprise software, it is optimized for large model training, post-training, and test-time inference.

In Stock
Inventory:99item
SKU:NVIDIA-DGX-GB300

Core Specifications

Number of GPUs
72 Blackwell Ultra
CPU
36 Grace
FP4 compute
1440 PFLOPS
GPU VRAM
20 TB
Shape
Rack-level liquid cooling

FullSpecifications

Number of GPUs72 Blackwell Ultra
CPU36 Grace
FP4 compute1440 PFLOPS
GPU VRAM20 TB
ShapeRack-level liquid cooling

* Sparse mode specifications: Dense performance is half that of sparse specs. Specs sourced from NVIDIA's official product page; actual configuration subject to factory settings.

Product BrochureDownload

Below isNVIDIA DGX GB300Product Brochure Preview: Click the button to generate and download a PDF version for internal review and selection archiving.

Huaying Technology Co., Ltd. · Product Brochure
NVIDIA DGX GB300
AI Computing / Data Centers· Supply from NVIDIA Authorized Distributors
SKU:NVIDIA-DGX-GB300
NVIDIA DGX GB300

The NVIDIA DGX GB300 is a rack-scale liquid-cooled AI system built on the NVIDIA Grace Blackwell Ultra superchip. It integrates 72 Blackwell Ultra GPUs and 36 Grace CPUs, delivering 2,592 Arm Neoverse V2 cores, up to 20TB of GPU memory (with bandwidth up to 576TB/s), and 37TB of total fast memory. With FP4 Tensor Core performance reaching 1440 PFLOPS (sparse), it offers up to 1.5x denser FP4 performance and 2x faster attention mechanisms compared to the DGX GB200. Designed with the NVIDIA MGX chassis, the DGX GB300 can scale to thousands of nodes when combined with DGX SuperPOD. Preloaded with NVIDIA Mission Control and AI Enterprise software, it is optimized for large model training, post-training, and test-time inference.

Number of GPUs72 Blackwell Ultra
CPU36 Grace
FP4 compute1440 PFLOPS
GPU VRAM20 TB
ShapeRack-level liquid cooling
Customer Service Hotline: 4006068166Main: 010-82981888
Building 7, Courtyard 2, No. 1 Guanghua Road, Chaoyang District, Beijing · huayingtech.com
NVIDIA DGX GB300Product Brochure
PDF · A4 · Includes full specifications and contact details

Brochures are generated in real-time from page content. If your browser blocks the download, allow pop-ups and try again.

Get selection and quote support
Customer Service Hotline 4006068166
Building 7, Compound 2, Guanghua Road, Chaoyang District, Beijing
Online Inquiry

Purchase Intent

Submit your requirements and contact details. A sales consultant will confirm pricing, stock availability, and delivery options with you. Product prices are subject to quotation.

Your information will be used solely for this selection and quotation discussion and for no other purpose.

Beijing ICP License No. 2026026760