NVIDIA GB200 NVL576

576× B200
2025
FP64
21.3
PFLOPS
FP32
43.2
PFLOPS
Power
960.0
kW Total
Memory
103680
GB Total
FP64
21.31
PFLOPS
27.0 TFLOPS/kW
FP32
43.20
PFLOPS
54.0 TFLOPS/kW
TF32
633.60
PFLOPS
1500.0 TFLOPS/kW
FP16
1296.00
PFLOPS
2700.0 TFLOPS/kW

System Details

GPU Configuration

GPU Count: 576 GPUs
Architecture: Blackwell
Interconnect:

System Specifications

Form Factor:
Total Power: 960.0 kW
Total Memory: 103680 GB
Memory Bandwidth: 4435200 GB/s

Precision Performance Breakdown

PrecisionSystem PerformancePer GPUEfficiency
FP64 21.312 PFLOPS 37.0 TFLOPS 27.0 TFLOPS/kW
FP32 43.200 PFLOPS 75.0 TFLOPS 54.0 TFLOPS/kW
TF32 633.600 PFLOPS 1100.0 TFLOPS 1500.0 TFLOPS/kW
FP16 1296.000 PFLOPS 2250.0 TFLOPS 2700.0 TFLOPS/kW
BF16 1296.000 PFLOPS 2250.0 TFLOPS 2700.0 TFLOPS/kW
FP8 2592.000 PFLOPS 4500.0 TFLOPS 5400.0 TFLOPS/kW
FP4 5184.000 PFLOPS 9000.0 TFLOPS 10800.0 TFLOPS/kW
INT8 2592.000 PFLOPS 4500.0 TFLOPS 5400.0 TFLOPS/kW

Powered by NVIDIA B200

This system utilizes 576 × 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 NVL576 with 576× B200 GPUs is designed for enterprise-scale workloads requiring 21.3 PFLOPS of compute power with high-speed interconnect for efficient multi-GPU communication.