NVIDIA DGX B300
- Form Factor 10U Rackmount
- Supported CPU Intel Xeon 6
- CPU Sockets 2
-
Max Supported Memory
4 TB ECC Supported
- Features 144 petaFLOPS inference
- 2.1TB total GPU memory
- 72 petaFLOPS training
- 8 NVIDIA Blackwell Ultra GPUs
NVIDIA DGX B300 is an enterprise AI platform designed to accelerate training, inference, and advanced AI reasoning. Featuring eight NVIDIA Blackwell Ultra GPUs connected through NVIDIA® NVLink™ and 2.1TB of GPU memory, it combines high computational performance with substantial memory capacity for demanding generative AI workloads. Compared with DGX B200, it delivers 1.5X the dense FP4 performance and 2X the attention performance. With a redesigned chassis compatible with NVIDIA MGX™ and traditional enterprise racks, DGX B300 gives businesses a flexible foundation for developing and deploying their next generation of AI applications.
Request a quote today
Share your email below to express your interest, and a Nextron representative will reach out shortly with detailed information and a personalized quote for your selected product.
More Specifications
- Configurable Configurable by Request
- Features 15.1kW power consumption (PSU) / 14.5kW (busbar)
- 8x OSFP ports with 8x ConnectX-8 VPI
- NVIDIA AI Enterprise, NVIDIA Mission Control, DGX OS / Ubuntu / Red Hat Enterprise Linux / Rocky Linux
- DIMM Type