Interactive 3D Model

Loading viewer...

Drag to rotate ┬╖ Scroll to zoom

ARCHITECTURE

Blackwell

SERIES

B200

Memory

384GB HBM3e

Bandwidth

10 TB/s

TDP

1000W

Pre-OrderB200┬╖Best for: Flagship AI

NVIDIA B200 Tensor Core GPU

Blackwell

|Data Center GPUs

Overview

The NVIDIA B200 Tensor Core GPU represents the pinnacle of AI computing with the new Blackwell architecture. Featuring 384GB of HBM3e memory and 10 TB/s bandwidth, the B200 can train trillion-parameter models with unprecedented efficiency.

Key Features

Blackwell Architecture
384GB HBM3e
NVLink 5.0
FP8/FP4

Ideal Use Cases

ai traininghpcdata analytics

Full Specifications

Complete technical specifications for the B200.

Memory

Memory: 384GB HBM3e
Memory Bandwidth: 10 TB/s

Compute

Tensor Cores: —
CUDA Cores: —
FP8 TFLOPS: 4500
FP16 TFLOPS: 2250
TF32 TFLOPS: 1125
FP64 TFLOPS: —

Interconnect

Interconnect: NVLink 5.0, 1.8TB/s

Physical

TDP: 1000W
Form Factor: SXM6
Cooling: Liquid Cooled

Manufacturing

Launch Date: H2 2024
Manufacturing Process: 4nm

Ready to deploy the B200?

Get a personalized quote within 24 hours. Our NVIDIA-certified engineers will help you architect the optimal configuration for your workload.

Related Chips

More chips in Data Center GPUs.

H100

In Stock

NVIDIA H100 Tensor Core GPU

Enterprise AI training and inference GPU with Transformer Engine and FP8 precision.

ArchitectureHopper

Memory80GB HBM3

TDP700W

H200

In Stock

NVIDIA H200 Tensor Core GPU

Next-gen AI GPU with 141GB HBM3e memory for larger models and faster inference.

ArchitectureHopper

Memory141GB HBM3e

TDP700W

GH200

On Order

NVIDIA GH200 Grace Hopper

Superchip combining Grace CPU with Hopper GPU for HPC and AI workloads.

ArchitectureHopper

Memory96GB HBM3 + 480GB LPDDR5X

TDP—

A100

In Stock

NVIDIA A100 Tensor Core GPU

Proven enterprise GPU for AI, HPC, and data analytics workloads.

ArchitectureAmpere

Memory80GB HBM2e

TDP400W