NVIDIA's new DGX Station 320G: 4 x Ampere GPUs + 320GB HBM2e for $150K

NVIDIA has just announced its super-beast new DGX Station 320G which packs the latest and greatest NVIDIA A100 Tensor Core GPUs which in this 4-way GPU system pack an insane 320GB of super-fast HBM2e memory.

This isn't a system for gamers but for the much higher-end AI and HPC (high-performance computing) markets, which will super-speed machine learning and data science performance where it's needed. Think about laboratories, home offices, VFX studios, research facilities and everything in between.

NVIDIA's new DGX Station 320G is the fastest-ever server-in-a-box designed for AI research.

10

Inside, we have 4 x NVIDIA A100 Tensor Core GPUs with a total of 320GB of HBM2e memory (80GB HBM2e memory per GPU) with multi-instance GPU (MIG) and third-gen NVLink connectivity. We have 200GB/sec of bi-directional bandwidth between GPU pairs, which is nearly 3x higher than the bandwidth available over the PCIe 4.0 standard.

We have an AMD EPYC CPU with 64 cores and 128 threads, up to 512GB of system memory, a 1.92TB NVMe M.2 SSD for the OS and up to 7.68TV of NVMe U.2 SSD for data cache.

The NVIDIA DGX Station 320G is a meticulously built AI server, and even better with its cooling -- check it out below:

10

The entire system has a refrigerant cooling system, which is absolutely beautiful, with 2 x 10GbE network connections, 4 x mini DisplayPort for video output. Better yet, the entire system is pretty much silent at just 37dB and is maintenance free, while the DGX Station 320G beast is powered by an equally huge 1500W PSU.

• BMW Group Production is using NVIDIA DGX Stations to explore insights faster as they develop and deploy AI models that improve operations.
• DFKI, the German Research Center for Artificial Intelligence, is using DGX Station to build models that tackle critical challenges for society and industry, including computer vision systems that help emergency services respond rapidly to natural disasters.
• Lockheed Martin is using DGX Station to develop AI models that use sensor data and service logs to predict the need for maintenance to improve manufacturing uptime, increase safety for workers, and reduce operational costs.
• NTT Docomo, Japan's leading mobile operator with over 79 million subscribers, uses DGX Station to develop innovative AI-driven services such as its image recognition solution.
• Pacific Northwest National Laboratory is using NVIDIA DGX Stations to conduct federally funded research in support of national security. Focused on technological innovation in energy resiliency and national security, PNNL is a leading U.S. HPC center for scientific discovery, energy resilience, chemistry, Earth science, and data analytics.

10

10

NVIDIA is charging $149,000 for its new DGX Station 320G, or you can get it on a monthly subscription for $9000 per month.