The NVIDIA DGX B100 System is a next-generation enterprise AI compute platform built around NVIDIA Blackwell (B100) technology to accelerate Generative AI, large-scale model training, fine-tuning, and high-throughput inference. Commonly described as the x86 / air-cooled Blackwell DGX class, DGX B100 is positioned for organizations that want a DGX-style, fully integrated system designed to bring Blackwell performance into modern data centers and scale into larger AI factories.
Key Highlights
- Blackwell-class DGX platform for GenAI: Designed to power enterprise AI pipelines—from training and fine-tuning to deployment.
- 8-GPU DGX system class (B100/HGX family): Referenced as part of the Blackwell DGX family in industry coverage (configuration can vary by SKU/OEM).
- High-speed GPU fabric (NVLink/NVSwitch family): DGX Blackwell systems are built to maximize multi-GPU throughput using NVIDIA’s NVLink/NVSwitch interconnect approach for fast GPU-to-GPU communication.
- Built to scale into DGX infrastructure: DGX systems are designed as building blocks for larger DGX platforms/SuperPOD-style deployments and AI factory architectures.
- DGX software + enterprise ecosystem: DGX systems are typically delivered as an integrated platform with NVIDIA’s software stack and enterprise support options (varies by offering/region).
