NVIDIA GB200 NVL36

36× B200
2025
FP64
1.3
PFLOPS
FP32
2.7
PFLOPS
Power
60.0
kW Total
Memory
6480
GB Total
FP64
1.33
PFLOPS
27.0 TFLOPS/kW
FP32
2.70
PFLOPS
54.0 TFLOPS/kW
TF32
39.60
PFLOPS
1500.0 TFLOPS/kW
FP16
81.00
PFLOPS
2700.0 TFLOPS/kW

System Details

GPU Configuration

GPU Count: 36 GPUs
Architecture: Blackwell
Interconnect:

System Specifications

Form Factor:
Total Power: 60.0 kW
Total Memory: 6480 GB
Memory Bandwidth: 277200 GB/s

Precision Performance Breakdown

PrecisionSystem PerformancePer GPUEfficiency
FP64 1.332 PFLOPS 37.0 TFLOPS 27.0 TFLOPS/kW
FP32 2.700 PFLOPS 75.0 TFLOPS 54.0 TFLOPS/kW
TF32 39.600 PFLOPS 1100.0 TFLOPS 1500.0 TFLOPS/kW
FP16 81.000 PFLOPS 2250.0 TFLOPS 2700.0 TFLOPS/kW
BF16 81.000 PFLOPS 2250.0 TFLOPS 2700.0 TFLOPS/kW
FP8 162.000 PFLOPS 4500.0 TFLOPS 5400.0 TFLOPS/kW
FP4 324.000 PFLOPS 9000.0 TFLOPS 10800.0 TFLOPS/kW
INT8 162.000 PFLOPS 4500.0 TFLOPS 5400.0 TFLOPS/kW

Powered by NVIDIA B200

This system utilizes 36 × NVIDIA B200 SXM 180GB GPUs, each delivering exceptional performance for AI and HPC workloads.

Per GPU TDP

1000W

Per GPU Memory

180 GB

Process Node

4nm

Architecture

Blackwell

Documentation & Resources

workstation-datasheet-dgx-spark-gtc25-spring-nvidia-us-3716899-web.pdf

NVIDIA • 2025-10-16

View Document ↗

NVIDIA DGX Documentation

System guides, deployment resources

Browse ↗

Typical Use Cases

AI/ML Training
High-Performance Computing
Data Analytics

The NVIDIA GB200 NVL36 with 36× B200 GPUs is designed for enterprise-scale workloads requiring 1.3 PFLOPS of compute power with high-speed interconnect for efficient multi-GPU communication.