Unlock Ultimate Computational Power, pre-reserve the all new NVIDIA H200s AI SuperCluster for $1.89/hr. Click to Know More.

Reserve the NVIDIA B200 at $2.99/hr for steady throughput, lower TCO, and reliable production workloads. Click to Reserve Now.

Pre-book the NVIDIA B300 at $3.99/hr for faster renders and heavy batch training with predictable performance. Click to Pre-Reserve B300 Now.

NVIDIA B300

Driving Enterprise AI Innovation

With NVIDIA Blackwell Ultra GPUs, DGX B300 delivers 11x faster inference and 4x quicker training compared to previous generations—all in a compact form factor that fits seamlessly

circle
circle

Features

An AI Factory for the Era of Generative Intelligence

  • LLM Inference & Training: Powered by NVIDIA Blackwell Ultra GPUs, DGX B300 accelerates large language model inference and training. Achieve 144 petaFLOPS of inference performance that enables your enterprise to operate at hyperscaler levels.
  • Efficient & Sustainable Innovation: With multiple power options and an industry-leading liquid cooling system, DGX B300 is engineered for unmatched energy efficiency, reducing costs and environmental impact.
NVIDIA DGX B300

Accelerate Your AI Journey

  • Revolutionary AI Infrastructure: Delivers enterprise-grade performance for demanding training and inference workloads.
  • Breakthrough Hardware: Powered by NVIDIA Blackwell Ultra GPUs and dual Intel Xeon processors for reliable, high-speed AI computing.
  • Unmatched Performance: Achieves 72 PFLOPS for FP8 training and 144 PFLOPS for FP4 inference, accelerating model development.
  • Scalable & Flexible: Designed for both bare metal and cloud GPU deployments to seamlessly scale with your business needs.
  • Optimized Connectivity: Advanced NVLink and high-speed networking ensure efficient data transfer and storage for large-scale AI applications.
Pre-Reserve Now button arrow arrow blink
Other GPU's Available:

H200

MI300x

H100

B100

RTX 3090

NVIDIA DGX B300 Specifications:

Spec Details
GPUs NVIDIA Blackwell Ultra GPUs
CPU Dual Intel® Xeon® Processors
Total GPU Memory 2.3TB
Performance 72 PFLOPS FP8 training; 144 PFLOPS FP4 inference
Networking 8x OSFP ports; Up to 800Gb/s NVIDIA InfiniBand/Ethernet; 2x dual-port QSFP112 NVIDIA BlueField-3 DPU
Storage OS: 2x 1.9TB NVMe M.2; Internal: 8x 3.84TB NVMe E1.S

NVIDIA DGX B300 vs. DGX B200 – Next-Level AI Performance

Feature NVIDIA DGX B300 NVIDIA DGX B200
GPUs NVIDIA Blackwell Ultra GPUs 8x NVIDIA Blackwell GPUs
GPU Memory 2.3TB total 1.44TB total
Performance 72 PFLOPS FP8 training, 144 PFLOPS FP4 inference 72 PFLOPS training, 144 PFLOPS inference
CPU Dual Intel Xeon Processors 2x Intel Xeon Platinum 8570 (112 cores, 4GHz Max)
Networking Up to 800Gb/s InfiniBand/Ethernet Up to 400Gb/s InfiniBand/Ethernet
Storage 8x 3.84TB NVMe E1.S internal storage 8x 3.84TB NVMe U.2 internal storage

Experience unparalleled AI performance with the NVIDIA DGX B300. Engineered with breakthrough Blackwell Ultra GPUs and dual Intel Xeon Processors, the DGX B300 delivers robust training and inference capabilities in a compact, scalable solution designed for modern data centers.