AI Supercomputing Platform
Massive datasets, complex simulations, and sophisticated models require multiple GPUs with high-speed interconnections to efficiently process and analyze the information in real-time, ensuring high performance and scalability across simultaneous AI jobs at every scale. By integrating 8 NVIDIA Hopper/Blackwell Tensor Core GPUs, the NVIDIA HGX™ AI supercomputing platform pack the power of NVIDIA GPUs, NVIDIA NVLink™, NVIDIA networking together to deliver exceptional compute performance for AI breakthroughs across industries.
Platform Overview
NVIDIA HGX™ B300
NVIDIA HGX™ B300 is built for the age of AI reasoning with enhanced compute and increased memory. Featuring 7X more AI compute than Hopper platforms, 2.3TB of HBM3e memory, and high-performance networking integration with NVIDIA ConnectX-8 SuperNICs, Blackwell Ultra delivers breakthrough performance on the most complex workloads from agentic systems and reasoning, to real time video generation for every data center.
Featured Product
Harnessing the power of NVIDIA Blackwell Ultra GPUs, the Ingrasys HGX B300 system unites advanced compute, HBM3e memory, and CX8 networking integration to supercharge AI reasoning. Featuring high-density storage and high expansion capability, the system can pair with an x86-based system of your choice to form a unified and powerful AI solution, paving the way for what’s possible in the next wave of innovation.
GB12580N
- GPU: NVIDIA HGX™ B300 (8 x NVIDIA Blackwell Ultra)
- Drives: 32 x U.2 NVMe
- Expansion Slots: 20 x PCIe 5.0 x16 FHHL Slots
- System Cooling: 30x 8080 Fans (N+1 Redundant)
- Power Supply: 6+6 3000W Power Supplies
- Form Factor: 12 U air-cooled System
