NVIDIA DGX GB300 NVL72
rack
72× B300
NVLink + NVSwitch
2025
FP32
6.0
PFLOPS
TF32
90.0
PFLOPS
Power
—
kW Total
Memory
21234
GB Total
36 GB300 superchips (72 B300 GPUs + 36 Grace CPUs) in liquid-cooled rack
FP32
6.00
PFLOPS
TF32
90.00
PFLOPS
FP16
180.00
PFLOPS
BF16
180.00
PFLOPS
System Details
GPU Configuration
GPU Model: NVIDIA
B300
SXM 288GB
GPU Count: 72 GPUs
Architecture: Blackwell Ultra
Interconnect: NVLink + NVSwitch
System Specifications
Form Factor: rack
Total Power: —
Total Memory: 21234 GB
Memory Bandwidth: 576000 GB/s
Precision Performance Breakdown
| Precision | System Performance | Per GPU | Efficiency |
|---|---|---|---|
| FP32 | 5.998 PFLOPS | 83.3 TFLOPS | — |
| TF32 | 90.000 PFLOPS | 1250.0 TFLOPS | — |
| FP16 | 180.000 PFLOPS | 2500.0 TFLOPS | — |
| BF16 | 180.000 PFLOPS | 2500.0 TFLOPS | — |
| FP8 | 360.000 PFLOPS | 5000.0 TFLOPS | — |
| FP4 | 1080.000 PFLOPS | 15000.0 TFLOPS | — |
Powered by NVIDIA B300
This system utilizes 72 × NVIDIA B300 SXM 288GB GPUs, each delivering exceptional performance for AI and HPC workloads.
Per GPU TDP
1400W
Per GPU Memory
288 GB
Process Node
4nm
Architecture
Blackwell Ultra
Documentation & Resources
workstation-datasheet-dgx-spark-gtc25-spring-nvidia-us-3716899-web.pdf
NVIDIA • 2025-10-16
NVIDIA DGX Documentation
System guides, deployment resources
Typical Use Cases
AI/ML Training
High-Performance Computing
Data Analytics
The NVIDIA DGX GB300 NVL72 with 72× B300 GPUs is designed for enterprise-scale workloads requiring 6.0 PFLOPS of compute power with NVLink + NVSwitch interconnect for efficient multi-GPU communication.