NVIDIA B300

The NVIDIA B300 represents the next evolution in the Blackwell architecture, delivering enhanced performance for the most demanding AI workloads. Built for training and deploying large-scale AI models, B300 offers improved efficiency and performance over its predecessors. With advanced NVLink interconnect and optimized memory bandwidth, it is ideal for LLMs, recommender systems, and real-time inference applications requiring maximum computational power.

NVIDIA B300 SXM6 AC

GPUs
8x NVIDIA B300 SXM6 AC
CPUs
2x Intel(R) Xeon(R) 6767P | 128 cores
Memory
3.07 TB
NVMe Total
32.65 TB
NVMe Disks
10
Location
Colorado

On demand

$5.50per card-hour

Interruptible

$3.95per card-hour

NVIDIA B300 SXM6 AC

GPUs
8x NVIDIA B300 SXM6 AC
CPUs
2x Intel(R) Xeon(R) 6767P | 128 cores
Memory
3.07 TB
NVMe Total
32.65 TB
NVMe Disks
10
Location
Colorado

On demand

$5.50per card-hour

Interruptible

$3.95per card-hour